In this paper, we first improve the influence propagation model by combining sentiment analysis and propose a new algorithm for node activation probability. Sentiment shift is the dynamic change of node attributes during the propagation of movie and television works. The algorithm for node activation probability has been further improved, and a model for influence propagation applicable to dynamic networks has been proposed. Finally, based on the data of the YouTube Internet platform, this paper analyzes the awareness of overseas audiences on Chinese film and television types, popular elements, publicity methods and domestic and foreign box office from 2010 to 2019. The findings indicate that 0.7379, 0.7031, and 0.5596 foreign viewers believe that three Chinese film and television genres, palace fighting, romance, and youth, are the most popular. Although there are still more than 0.42 million overseas viewers who prefer martial arts genres, they are in a downward stage compared to before. In terms of publicity, more than 0.91 percent of them believe that online social networking areas and chat groups for movie and TV enthusiasts are the most effective forms of publicity. Therefore, this study is an important reference for the development of communication strategies.