With the rapid development of society, online comment has become an important reference for consumers to make purchase decisions in the film industry, and has an important influence on the box office. Based on the sentiment analysis method, the sentiment information in the film comment texts is extracted and quantified, and the box office prediction model is established to study the influence of the number of comments, the sentiment value of comments, and the early box office revenue on total box office revenue during the movie schedule. The empirical analysis of domestic films released in the Chinese mainland from 2016 to 2019 shows that the box office prediction model, established by independent variables including the box office revenue of the previous week, the number and sentiment value of online comments published on the previous day, can effectively predict the total box office revenue. The model can also be used to forecast sales of other products after improvement.