In recent years, research on deep learning has become an important topic in artificial intelligence, has achieved excellent results in several evaluation fields, and has become a cutting-edge method for solving various evaluation problems. For this reason, this paper explores the effects of China’s international music communication from three dimensions: cognition, attitude, and behavior, and establishes an evaluation system containing 11 secondary indicators. Then, an assessment model of the music communication effect is constructed using the deep learning network method. The constructed assessment model was used to evaluate the impact of international communication on Chinese music. According to the output of the model, it was determined that the communication power of the selected 1,120 Chinese music songs belonged to the solid and average levels, accounting for 71.34% of the total, and the comprehensive evaluation score was between 2.5 and 4.3. This indicates that Chinese music has achieved better dissemination effectiveness and influence with the help of various music platforms.