In short

Every plane in three-dimensional space has the equation ax + by + cz = d, where (a, b, c) is a normal vector. This article gives you four recipes for writing that equation when you are told different geometric facts: a point and a normal, a parallel plane and a point, a line and a point, or two parallel lines. Each recipe is just a variation on the same core idea \u2014 the normal vector decides everything.

Imagine you are designing the roof of a temple. The roof is a flat surface \u2014 a plane \u2014 that must pass through three fixed pillars at known positions. Or perhaps you need a wall that runs parallel to an existing wall but passes through a specific corner. Or a ramp that contains a given railing (a line) and also touches a particular point on the ground.

All of these are the same mathematical question: find the equation of a plane, given geometric constraints.

You already know the raw equation of a plane: ax + by + cz = d, where (a, b, c) is a vector perpendicular to the plane. That fact alone \u2014 one equation, one normal \u2014 is enough to describe every flat surface in three-dimensional space.

But in practice, problems never hand you the equation directly. Instead, they give you geometric clues. "Find the plane through (1, 2, 3) with normal \hat{i} - 2\hat{j} + \hat{k}." Or: "Find the plane through (4, 0, 1) parallel to 2x - y + 3z = 7." Or something stranger: "Find the plane that contains the line \dfrac{x-1}{2} = \dfrac{y+3}{1} = \dfrac{z}{-1} and also passes through the point (0, 5, 2)."

Each of these is a puzzle, and each has a clean solution that flows from one principle: to write the equation of a plane, find its normal vector, then use any point on the plane to pin down the constant.

Why the normal? Because the normal vector (a, b, c) is the plane's identity card. Two planes with the same normal are parallel. A plane with a different normal tilts at a different angle. Once you know the tilt (the normal) and one point that the plane passes through, the plane is completely determined \u2014 there is exactly one flat surface with that tilt passing through that point.

Form 1: Plane through a point with a given normal

This is the most direct case. You know a point P_0 = (x_0, y_0, z_0) on the plane, and you know a normal vector \vec{n} = (a, b, c) perpendicular to the plane.

Any other point P = (x, y, z) lies on the plane if and only if the vector from P_0 to P is perpendicular to \vec{n}. Two vectors are perpendicular when their dot product is zero:

\vec{n} \cdot (\vec{P} - \vec{P_0}) = 0

Expanding:

a(x - x_0) + b(y - y_0) + c(z - z_0) = 0

That is the equation. The normal vector gives the coefficients a, b, c, and the known point gives the constant term.

Plane through a point with a given normal

If the plane passes through (x_0, y_0, z_0) and has normal vector \vec{n} = (a, b, c), its equation is

a(x - x_0) + b(y - y_0) + c(z - z_0) = 0

or equivalently, ax + by + cz = d where d = ax_0 + by_0 + cz_0.

The vector form says the same thing more compactly. If \vec{r} is the position vector of a general point on the plane and \vec{r_0} is the position vector of the known point:

\vec{n} \cdot (\vec{r} - \vec{r_0}) = 0 \quad \Longleftrightarrow \quad \vec{n} \cdot \vec{r} = \vec{n} \cdot \vec{r_0}
Plane through a point with a given normal vectorA 3D perspective view of a plane. A point P0 is marked on the plane. A vector n points straight up from the plane, perpendicular to it. A second point P is on the plane, and the vector from P0 to P lies flat in the plane, perpendicular to n. n\u20d7 P\u2080 P P \u2212 P\u2080 n\u20d7 \u00b7 (P \u2212 P\u2080) = 0
A point $P$ lies on the plane if and only if the vector from $P_0$ to $P$ is perpendicular to the normal $\vec{n}$. The dot product captures this perpendicularity condition exactly.

Form 2: Plane parallel to a given plane

Suppose you want a plane that is parallel to 2x - y + 3z = 7 and passes through (1, 4, -2).

Parallel planes have the same normal direction. The given plane has normal (2, -1, 3), so the new plane does too. Its equation is therefore

2x - y + 3z = d

for some constant d. To find d, substitute the known point:

2(1) - (4) + 3(-2) = 2 - 4 - 6 = -8

So d = -8, and the plane is 2x - y + 3z = -8.

Plane parallel to a given plane

A plane parallel to ax + by + cz = d_1 and passing through (x_0, y_0, z_0) has equation

ax + by + cz = ax_0 + by_0 + cz_0

The coefficients stay the same; only the constant changes.

The logic is almost trivially short \u2014 and that is the point. Parallel planes share a normal, so the only degree of freedom left is how far along that normal the plane sits. The known point fixes that distance.

Parallel planes sharing the same normal vectorTwo parallel planes shown in perspective. Both have the same normal vector n pointing upward. A point P is marked on the lower plane. The only difference between the two planes is the constant d in the equation. given plane: ax + by + cz = d\u2081 new plane: ax + by + cz = d\u2082 n\u20d7 P (given point)
Two parallel planes have the same normal $\vec{n}$ but different constants $d_1$ and $d_2$. To find the new plane, keep the normal from the given plane and substitute the known point to find the new constant.

Form 3: Plane through a line and a point

Now the geometry becomes more interesting. You are given a line and a point not on the line, and you need the unique plane that contains both.

Consider the line \dfrac{x - 1}{2} = \dfrac{y + 3}{1} = \dfrac{z}{-1} and the external point Q = (0, 5, 2). The line passes through A = (1, -3, 0) in the direction \vec{d} = (2, 1, -1).

To write the plane's equation, you need its normal vector. What do you know? The normal must be perpendicular to the direction of the line \vec{d} (because the line lies in the plane). And the normal must be perpendicular to the vector \vec{AQ} (because this vector also lies in the plane \u2014 both A and Q are in the plane).

A vector perpendicular to two given vectors is their cross product:

\vec{n} = \vec{d} \times \vec{AQ}

Compute \vec{AQ} = Q - A = (0 - 1, \; 5 - (-3), \; 2 - 0) = (-1, 8, 2).

\vec{n} = \vec{d} \times \vec{AQ} = \begin{vmatrix} \hat{i} & \hat{j} & \hat{k} \\ 2 & 1 & -1 \\ -1 & 8 & 2 \end{vmatrix}
= \hat{i}(1 \cdot 2 - (-1) \cdot 8) - \hat{j}(2 \cdot 2 - (-1)(-1)) + \hat{k}(2 \cdot 8 - 1 \cdot (-1))
= \hat{i}(2 + 8) - \hat{j}(4 - 1) + \hat{k}(16 + 1)
= 10\hat{i} - 3\hat{j} + 17\hat{k}

So \vec{n} = (10, -3, 17). The plane passes through A = (1, -3, 0):

10(x - 1) - 3(y + 3) + 17(z - 0) = 0
10x - 3y + 17z = 19

You can verify: the point Q = (0, 5, 2) gives 10(0) - 3(5) + 17(2) = -15 + 34 = 19. Correct.

Plane through a line and a point

Given a line through A with direction \vec{d}, and a point Q not on the line:

  1. Compute \vec{AQ} = Q - A.
  2. Compute the normal: \vec{n} = \vec{d} \times \vec{AQ}.
  3. Use Form 1 with point A and normal \vec{n}.

The cross product produces the unique direction perpendicular to both vectors lying in the plane.

Plane determined by a line and an external pointA perspective view showing a plane determined by a line and a point Q not on the line. The line passes through point A with direction vector d. The vector AQ joins A to Q. The normal n, computed as d cross AQ, points upward from the plane. line A d\u20d7 Q AQ\u20d7 n\u20d7 = d\u20d7 \u00d7 AQ\u20d7
The plane through a line (with direction $\vec{d}$) and an external point $Q$. Both $\vec{d}$ and $\vec{AQ}$ lie in the plane. Their cross product gives the normal $\vec{n}$, the one direction perpendicular to the plane.

Form 4: Plane through two parallel lines

Two parallel lines that are not the same line define a unique plane. Think of two parallel railway tracks running across a flat field — there is exactly one flat surface that contains both tracks.

The setup: line \ell_1 passes through A with direction \vec{d}, and line \ell_2 passes through B with the same direction \vec{d} (they are parallel, so they share a direction vector).

The strategy is identical to Form 3. Both \vec{d} and \vec{AB} lie in the plane, so the normal is:

\vec{n} = \vec{d} \times \vec{AB}

Then use any known point (A or B) with this normal to write the equation.

Here is a quick concrete example. Take \ell_1 : \dfrac{x}{1} = \dfrac{y - 1}{2} = \dfrac{z - 2}{1} and \ell_2 : \dfrac{x - 3}{1} = \dfrac{y}{2} = \dfrac{z + 1}{1}. Both have direction \vec{d} = (1, 2, 1). The first passes through A = (0, 1, 2) and the second through B = (3, 0, -1).

Compute \vec{AB} = (3, -1, -3). Then:

\vec{n} = \vec{d} \times \vec{AB} = \begin{vmatrix} \hat{i} & \hat{j} & \hat{k} \\ 1 & 2 & 1 \\ 3 & -1 & -3 \end{vmatrix} = \hat{i}(-6 + 1) - \hat{j}(-3 - 3) + \hat{k}(-1 - 6) = (-5, 6, -7)

Using point A = (0, 1, 2): -5(0) + 6(1) - 7(2) = -8. So the plane is -5x + 6y - 7z = -8, or equivalently 5x - 6y + 7z = 8.

Plane through two parallel lines

Given parallel lines through A and B, both with direction \vec{d}:

  1. Compute \vec{AB} = B - A.
  2. Compute the normal: \vec{n} = \vec{d} \times \vec{AB}.
  3. Use Form 1 with point A and normal \vec{n}.

This is the same recipe as Form 3 \u2014 the second line provides a point (B) not on the first line, and the cross product does the rest.

Plane through two parallel linesA perspective view showing two parallel lines lying in a plane. Both lines have the same direction vector d. The vector AB connects a point on the first line to a point on the second line. The normal n, computed as d cross AB, points upward from the plane. line 1 line 2 A B AB\u20d7 n\u20d7
Two parallel lines with the same direction $\vec{d}$ determine a unique plane. The vector $\vec{AB}$ connecting a point on each line lies in the plane, and $\vec{n} = \vec{d} \times \vec{AB}$ gives the normal.

Notice that if the two lines are the same (not just parallel but actually coincident), then \vec{AB} is parallel to \vec{d}, and the cross product \vec{d} \times \vec{AB} = \vec{0}. The normal vanishes \u2014 and that makes geometric sense: a single line does not determine a unique plane, because infinitely many planes contain the same line.

The pattern behind all four forms

Look at what you have done four times now. Each form asks: what is the normal? Once you know the normal and one point, you have the plane.

Given information How to get \vec{n} How to get d
Point + normal \vec{n} is given directly Dot \vec{n} with the given point
Point + parallel plane Copy \vec{n} from the parallel plane Dot \vec{n} with the new point
Line + external point \vec{n} = \vec{d} \times \vec{AQ} Dot \vec{n} with any point on the line
Two parallel lines \vec{n} = \vec{d} \times \vec{AB} Dot \vec{n} with any point on either line

The column "how to get d" is the same in every row: substitute a known point into ax + by + cz. The entire art is finding the normal.

Worked examples

Example 1: Plane through a point, parallel to a given plane

Problem. Find the equation of the plane through (3, -1, 2) parallel to x + 2y - z = 5.

Step 1. Read off the normal from the given plane.

\vec{n} = (1, 2, -1)

Why: in ax + by + cz = d, the coefficients (a, b, c) are the normal vector. Parallel planes share the same normal.

Step 2. Write the new plane with the same normal but unknown constant.

x + 2y - z = d

Why: a plane parallel to x + 2y - z = 5 has the same left side; only the right side changes.

Step 3. Substitute the known point (3, -1, 2) to find d.

3 + 2(-1) - 2 = 3 - 2 - 2 = -1

Why: the point must satisfy the equation, so plugging it in gives you d.

Step 4. Write the final equation.

x + 2y - z = -1

Result: x + 2y - z = -1.

A 2D cross-section view (setting $z = 0$): the original plane's trace is the dashed line $x + 2y = 5$, and the new plane's trace is the solid red line $x + 2y = -1$. Both lines have the same slope (same normal direction) but different intercepts \u2014 parallel, exactly as required. The point $(3, -1)$ sits on the red line.

The two planes are parallel \u2014 same normal direction, different distance from the origin. The new plane is pushed to a different position along the normal, landing exactly on the given point.

Example 2: Plane containing a line and a point

Problem. Find the equation of the plane that contains the line \dfrac{x - 2}{1} = \dfrac{y - 1}{-1} = \dfrac{z + 1}{3} and passes through (1, 0, 1).

Step 1. Extract the line's data.

The line passes through A = (2, 1, -1) with direction \vec{d} = (1, -1, 3).

Why: the denominators of the symmetric form give the direction ratios, and the numerators give the fixed point.

Step 2. Compute \vec{AQ} where Q = (1, 0, 1).

\vec{AQ} = (1 - 2, \; 0 - 1, \; 1 - (-1)) = (-1, -1, 2)

Why: you need a second vector lying in the plane. The vector from the line's point to the external point does that job.

Step 3. Compute the normal \vec{n} = \vec{d} \times \vec{AQ}.

\vec{n} = \begin{vmatrix} \hat{i} & \hat{j} & \hat{k} \\ 1 & -1 & 3 \\ -1 & -1 & 2 \end{vmatrix}
= \hat{i}((-1)(2) - (3)(-1)) - \hat{j}((1)(2) - (3)(-1)) + \hat{k}((1)(-1) - (-1)(-1))
= \hat{i}(-2 + 3) - \hat{j}(2 + 3) + \hat{k}(-1 - 1) = \hat{i}(1) - \hat{j}(5) + \hat{k}(-2)

So \vec{n} = (1, -5, -2).

Why: the cross product of two vectors in the plane gives the unique direction perpendicular to the plane.

Step 4. Use point A = (2, 1, -1) to write the equation.

1(x - 2) - 5(y - 1) - 2(z + 1) = 0
x - 2 - 5y + 5 - 2z - 2 = 0
x - 5y - 2z = -1

Step 5. Verify with the external point Q = (1, 0, 1): 1 - 0 - 2 = -1. Correct.

Result: x - 5y - 2z = -1.

A 2D cross-section (projecting onto the $xy$-plane by setting $z = 0$). The solid red line is the trace of the plane $x - 5y - 2z = -1$ on the $xy$-plane, which is $x - 5y = -1$. The projected direction of the line and the external point both land on this trace, confirming the plane passes through both.

Both the line and the point lie in the plane \u2014 and the cross product guaranteed it. The normal (1, -5, -2) is perpendicular to both the line's direction (1, -1, 3) and the bridge vector (-1, -1, 2). You can verify: \vec{n} \cdot \vec{d} = 1(1) + (-5)(-1) + (-2)(3) = 1 + 5 - 6 = 0.

Common confusions

A few traps that catch students in 3D geometry problems involving planes.

Going deeper

If you came here to learn the four standard forms and see them applied, you have it \u2014 you can stop here. The rest of this section explores what happens when the geometric data is under-determined or over-determined, and the parametric form that gives you every point on a plane without using the normal at all.

When the cross product vanishes

In Forms 3 and 4, the entire method depends on the cross product \vec{d} \times \vec{AQ} (or \vec{d} \times \vec{AB}) being nonzero. When is it zero? When the two vectors are parallel \u2014 meaning they point in the same (or opposite) direction.

For Form 3, \vec{d} \parallel \vec{AQ} means the external point Q lies on the line itself. That is not really an "external" point \u2014 it is on the line. And a single line does not determine a unique plane.

For Form 4, \vec{d} \parallel \vec{AB} means the two "parallel lines" are actually the same line. Again, no unique plane.

In both cases, the vanishing cross product is the algebra telling you that the geometric data is insufficient.

The parametric form of a plane

There is a completely different way to describe a plane. Instead of using a normal vector (the direction not in the plane), you can use two direction vectors that are in the plane.

If the plane contains the point \vec{r_0} and is spanned by two non-parallel vectors \vec{u} and \vec{v}, then every point on the plane can be written as:

\vec{r} = \vec{r_0} + s\,\vec{u} + t\,\vec{v}

where s and t are real numbers that range over all values. This is the parametric form of the plane \u2014 two parameters, one for each independent direction you can move in.

The parametric form is the plane analogue of the parametric form of a line (\vec{r} = \vec{a} + t\vec{d}, one parameter, one direction). A line has one degree of freedom; a plane has two.

To convert from parametric to Cartesian: compute \vec{n} = \vec{u} \times \vec{v} and then write \vec{n} \cdot (\vec{r} - \vec{r_0}) = 0. The normal is hiding inside the parametric description \u2014 it is the cross product of the two spanning vectors.

A general principle: the family of planes through a line

When a line lies in a plane, the line's direction vector \vec{d} is perpendicular to the normal: \vec{n} \cdot \vec{d} = 0. This gives one equation constraining the normal. But one equation in three unknowns (a, b, c) leaves a two-parameter family of solutions \u2014 which corresponds to the geometric fact that infinitely many planes pass through a single line.

Adding an external point supplies a second constraint (\vec{n} \cdot \vec{AQ} = 0), reducing the family to a single plane. Adding a second parallel line supplies the same kind of second constraint (\vec{n} \cdot \vec{AB} = 0).

This viewpoint \u2014 each piece of geometric data adds one constraint on the normal \u2014 is the unifying idea behind all four forms. The normal has three components, so you need enough constraints to pin it down (up to scaling). A point and a normal give you the normal directly (three constraints). A line and a point give two constraints. Two parallel lines give two constraints. In each case, the result is one plane.

Connection to the intercept form

When a plane has three non-zero intercepts \u2014 it crosses the x-axis at (a, 0, 0), the y-axis at (0, b, 0), and the z-axis at (0, 0, c) \u2014 the intercept form is

\frac{x}{a} + \frac{y}{b} + \frac{z}{c} = 1

This is just the three-point form applied to the three intercept points. It is a special case of "plane through three points," which itself is a double application of the cross-product recipe: take two vectors from one intercept point to the other two, cross them to get the normal, then write the equation. The intercept form packages the result into a single elegant fraction.

Where this leads next

You now have a toolbox for writing the equation of a plane given various geometric clues. The natural next steps: