Is the method of images applicable to gravity?

While rare, there are a few uses of the method of images to gravitational problems. As lurscher says, the problem is finding equipotential surfaces. In most problems, such a surface doesn't exist, and hence the scare use of the method of images in GR.

One class of problems for which it does applies are the so-called Dirichlet problems. Suppose one was interested in solving for the metric in some region, with specified boundary conditions on the boundary surface. This is not usually what is done--usually the entire spacetime is solved for. For the case of Dirichlet boundary conditions (requiring the metric to approach some specified value on the boundary surface), image charges can be useful. In this case the image charges could correspond to image black holes, for example.

However, this is somewhat of an exotic problem, and I've only seen perhaps one or two examples where image charges have been used.