The invention discloses a method and a system for identifying a signal modulation mode based on a convolutional neural network, and belongs to the technical field of signal detection and identification. The method comprises the following steps: adding noise to one of two paths of noiseless signals sent by a signal source; generating a high-order cumulant and a two-dimensional matrix as training labels, and generating a high-order cumulant and a two-dimensional matrix as data input quantities; obtaining a plurality of de-noising feature models, and generating an identification model; and obtaining a signal sent by the signal source, extracting I/Q information, truncating the high-order cumulant of the I/Q information, generating a two-dimensional matrix, sending the two-dimensional matrix into the identification model to carry out modulation identification on the signal, and outputting a signal modulation mode. According to the method, the generalization capability and the identification accuracy of the classifier are improved. The number of actually received signal samples is reduced, the influence of noise is effectively inhibited by using unsupervised de-noising self-coding, andthe accuracy of a final identification model is improved.