Abstract: In order to obtain the high precision automatic identification result of the intrusion behavior of optical communication system, an automatic identification technology of the intrusion behavior of optical communication system based on deep learning is proposed. Firstly, the fiber Bragg grating sensing technology is used to collect the intrusion behavior signal, and then the feature vector of the intrusion is extracted from the signal. Finally, the intrusion behavior recognition model of the optical communication system is established by using the deep learning network, and the simulation experiment is carried out. As a result, this method can automatically identify the intrusion behavior of high-precision optical communication system and reduce the error of automatic identification of intrusion behavior of optical communication system.