The people's income and living standards were improve in recent years. The number of tourist were increased year by year. Government must to plan and to satisfy people for tourism facilities and quality requirements. The forecasting of tourist number is an important job, it also were attended gradually. If it can predict passenger number, and then we will control the effective use of tourism resources in the future, as well as the promotion of tourism satisfaction tourists. In this study, ten prediction methods and time series analysis of ARIMA model were used to predict the tourist number. MAPE was used to compare the forecast accuracy. Finally, we can understand the changes of tourist number and can plan in the future.