Tourism seasonality refers to the distribution characteristics and trends of relative concentration of tourist flows over a period of time, which reflects the seasonal distribution and duration of tourist season. Based on the monthly data of tourist flow and online booking volume from 2008 to 2012, this paper analyzes the seasonal characteristics of tourist flow during the year, and points out the off-season and high-season. Then we combined with the climate data provided by the Weather Underground weather website to evaluate each month's climate comfort, and according to the degree of comfort to distinguish its different levels of comfort. After the climate comfort index and the holiday index are assigned, the relationship between the changes of tourist flow and online booking, climate comfort, festival factor was analyzed by OLS regression method. Finally, the three factors are recognized as the independent variables, the SVR method is used to forecast the monthly tourist flow. The results show that: 1) The annual variation of tourist flow and online booking volume showed the characteristics of “W” type; 2) The climate comfort and the length of holiday have obvious influence on the change of the tourist flow, also the quantity of online booking is positively correlated with the tourist flow; 3) SVR method is more accurate than multivariate linear regression method in predicting the monthly tourist flow. This study is helpful for the management personnel to predict the future tourist flow accurately and combine the carrying capacity to arrange the coping strategies, which has important guiding value for the scientific management.