An efficient algorithm is proposed for the radar cross-section (RCS) prediction of complex target with electronically large size, which is a combination of geometrical optics and physical optics (GO–PO) method. The method taking the multiple reflections into account is applied to the electromagnetic scattering analysis of a satellite model. Then RCS curves of entire satellite model and the model without antenna structure are figured out. Based on the simulated echoes, the traditional inverse synthetic aperture radar (ISAR) images are discussed. Moreover, an application of motion compensation technique based on the joint time-frequency analysis is presented for ISAR imaging of the moving satellite that has both translational and rotational movements. Numerical results show good performance of GO–PO method in accuracy and efficiency and the great influence of the antenna with corner structures on the scattering characteristic of the satellite.