First - know that a laser spot is not a "disk"; usually you will find that the intensity drops off from the middle to the edge, so "size of the spot" is an imprecise concept (where is the edge of a gradually decreasing curve?). Secondly, unless the laser is normally incident on the surface, the spot will most likely not be circular but elliptical; and unless you are viewing the surface normally, you will add an additional distortion.
All this makes it difficult to do this right. I would recommend that you place some graph paper (remember graph paper? It's what we used before computers could draw graphs... - it comes with markings in mm, with major markings for cm) on the surface where you project your laser dot, and take a picture of it. For precise measurements you can buy optical grids in many sizes - more expensive but much more accurate.
Unless the grid is perfectly normal to the axis of your camera, it will be distorted; but if you align it properly you will be able to count the pixels between lines on the grid in your image.
There are several tricks you can use to be more accurate. If you tilt the grid slightly (rotate about the normal) then the lines will cross rows and columns of pixels: this will allow you to find the place where the line aligns exactly with a pixel, and lets you get to sub-pixel accuracy. Second, you should count pixels across a number of lines; say you can see ten lines of the grid, you can calculate the distance by dividing the number of pixels between line 1 and 10 by 9 to get the number of pixels per mm.
Even more accurate would be to write down the pixel number for each line, then fit a straight line through the points. When you then subtract the best fit from the data you will be able to see whether there is any pincushion distortion (non linearity in the scaling of the camera with respect to position) at which point you can compute the scale factor at any point.
Finally, you could do this mathematically if you know the distance to the object exactly, and you know the focal length of the lens. This is actually quite hard to do because it's unlikely that you know the exact location of the optical center of the lens, but that depends on your setup.
In principle, if you have an object that is distance $o$ from the lens with focal length $f$, then in order for the object to be in focus you need the lens at a distance $d$ from the focal plane such that
$$\frac{1}{f} = \frac{1}{d}+\frac{1}{o}$$
Or, if you know the focal length of the lens but not the center of the lens, but you can measure the distance from the object to the sensor, $s = o + d$, then you can do
$$\begin{align}\frac{1}{f} &= \frac{1}{d} + \frac{1}{s - d}\\
&=\frac{s-d}{d(s-d)} + \frac{d}{d(s-d)}\\
&=\frac{s}{d(s-d)}\\
d(s-d)&=f\cdot s\end{align}$$
We can solve this quadratic for d:
$$d^2 - ds + fs = 0\\
d = \frac{s±\sqrt{s^2-4fs}}{2}$$
And then the magnification follows from
$$M = \frac{d}{o} = \frac{d}{s-d}$$
If you know the actual pixel size (usually this is given by the manufacturer) you can then get the magnification from this equation and get your pixel to mm conversion:
$$c = \mathrm{(Pixel\ size)\cdot M}$$.
So if your pixel is 0.1 mm and your magnification is 3, then one pixel corresponds to 0.3 mm in the image (at that particular distance).
