The invention discloses a facial expression recognition method based on auxiliary learning of a significant expression change area. According to the method, an auxiliary learning network is established to carry out feature extraction on a significant expression change area in a facial expression image; parameters of the first three feature extraction layers of the main network and the auxiliary learning network are shared, features extracted by the fourth feature extraction layer and the fifth feature extraction layer of the auxiliary learning network and features extracted by the fourth feature extraction layer and the fifth feature extraction layer of the main network are subjected to weighted fusion, and therefore the main network structure can learn features of some significant expression areas in the auxiliary network; processing the face expression data set by using a face detection and positioning algorithm to obtain a face region image for training a main network. The face region image is preprocessed to obtain an image with an expression significant change region, and the image is used for training an auxiliary learning network, so that an expression recognition main network can pay more attention to the expression significant change region, and expression features with higher recognition capability and robustness are extracted.