According to the gray distribution character of vignetting image photographed by interleaving assembled CCD camera using reflector, the method of two-dimensional quadrics fitting based on steepest descent method is introduced in this paper to solve the problem. The gray compensation algorithm based on two-dimensional quadrics fitting using matlab simulation is introduced after the analysis of the vignetting physical model of interleaving assembled CCD camera using reflector. Parameters of Hesse matrix are also deduced at the same time. The simulation result has illuminated that the parameters of irregular two-dimensional quadrics, such as vignetting image gray level, can be estimated effectively by this method. The actual processing result shows the vignetting phenomenon can be removed effectively, and consequently the quality of image is improved. We have used the new algorithm to restore 782×582 gray image on the test platform, the programme’s running time is 24 ms, so the algorithm can completely meet the practical need of the projects. Since no relative optical and geometrical parameters are needed, this method can be widely used in projects.