With micro-blog has gradually become the main platform for film promotion, the study of the relationship between micro-blog's data and box office is particularly important. By using the method of empirical research, based on the 3 platforms to collect a week before the release of the film and the release of the data during the period of each day. Making a correlation analysis of these factors and box office, and finding that each factor has a positive correlation with the box office. The regression was establish by Principal component regression analysis, and using data collected to validate it. The result shows that the model has good predictive performance and it can provide more practical value for the marketing strategy of the film producer.