Aiming at the problems of target detection in large scene and multi-class synthetic aperture radar (SAR) images, the target detection in SAR images based on deep learning is analyzed. This paper reviews the current status of target detection in SAR images from three aspects of multi-scale, complex background and large scene, summarizes the existing problems of target detection in SAR images, and looks forward to the future work. The results show that the study can provide a reference for related work.