Facial expression recognition technology, as an important branch in the field of computer vision, aims to identify and analyze human facial expressions. This technology infers human emotional states by analyzing facial features, dynamic changes in expressions, and details of expressions in specific areas. In this paper, the basic methods of facial expression recognition are first introduced, and the effectiveness of deep learning in this task is discussed. In the experimental part, four experiments are designed to explore the impact of different network factors on the accuracy of face recognition networks. Then, this paper designs a expression recognition system and implements the model. Users can upload facial expression pictures and get expression recognition results through the system. The system can recognize facial expressions in real time, and apply the theoretical results of the experiment to practical scenarios, which provides convenience for people's use.