In order to solve the problems encountered in the process of manually organizing photos nowadays, a deep learning-based automatic photo archiving system is designed and implemented. In the system, the deep learning-based face detection method MTCNN and face recognition method FaceNet are adopted, in which convolutional neural network is used to realize the detection, feature extraction and recognition of faces in photos. The automatic photo archiving system can provide an intuitive user interface and convenient operation, and users can realize the archiving management of photos through simple interaction. In the process of implementation, compared with the traditional face recognition method, the deep learning-based face recognition method will improve the accuracy of face recognition and help users organize photos more accurately.