Abstract：The paper proposes a particle swarm optimization and mutual information algorithm for UAV image registration. The purpose is to solve the problems of matching errors and low efficiency of traditional algorithms in multi-view remote sensing images. The algorithm simulates the deformation of building in multi-view sensing image via affine sampling. As a consequent, the image registration problem converts into an optimization problem for searching the best affine transformation. Therefore, This paper aims at analyze the definition of fitness function, the searching space and the combination of parameters. In the section of experiments, we carried out a series of experiments for 4 pairs of UAV remote sensing images. These experimental results show that the proposed algorithm can realize multi-view remote sensing image registration. This method is faster than the exhaustive mutual information search algorithm, and its accuracy is higher than that of the simplex method. In other words, the proposed method improves the robustness of different view image registration effectively.