Abstract: In recent years, face attribute editing has attracted widespread attention from the public and researchers. The major entertainment software has applied the face attribute editing technology to image editing, and realized functions such as changing human hair color, adding beard, and changing age. Researchers used face attribute editing algorithms to generate different postures, expressions, and face images under different lighting to assist the face recognition system and improve recognition accuracy. This paper mainly introduced the development process of face attribute editing based on deep learning, and classified and summarized the main technical routes and related algorithms. Face attribute editing was mainly composed of three parts: face attribute editing based on attribute label, face attribute editing based on reference condition information, and face attribute editing based on latent code. This paper summarized and prospected the development trend of this topic. It analyzed the development potential of higher-resolution face attribute editing and other aspects of potential room for improvement.