Image registration is a key technology for image-guided surgery, image fusion, organ map generation, tumor and bone growth monitoring and other clinical applications. It is also a very challenging problem. In recent years, deep learning technology has exerted an important influence on the research of medical image processing methods, and has developed rapidly in the field of medical image registration. Research on medical image registration using deep learning technology was reviewed. Firstly, according to the deep learning model, medical image registration methods were divided into three categories, including supervised, weakly supervised and unsupervised medical image registration. Then the research progress at home and abroad was introduced, and the advantages and disadvantages of these research methods were summarized. On this basis, the commonly used deep learning registration framework and evaluation criteria were described, and the commonly used open source medical image data sets were summarized. Finally, the existing problems of deep learning technology in the field of medical registration image were analyzed, and the future development direction was forecasted.