The research discusses the elements that create film value. We construct a forecasting model then compare actual and estimated revenues of box offices. The result can provide information as a reference to investors and relevant movie workers for investing and production references. We use a regression model to analyze the Taiwanese films data from 2002 to 2012. We found components such as "award winner (actor/actress)", "release timing", "series" and "award winner (directors)" have significant influences on the performance of box offices. The average estimated revenues of box office is 56.40% of the actual ones. The result can increase the predicting power of film revenues of box office. We suggest further study can address on the classification and setting standards for qualitative components, such as "shooting technique" and "type of story and ways of demonstration", then further decoding the film value.