Nowadays, with the rapid development of tourism, traveling has become a normal state. Moreover, in China, the number of tourists in many scenic spots is obviously seasonal, and the accurate prediction of tourists has become a concern. In this paper, the parameter optimal model of random forest is constructed, and the model parameters are adjusted by goodness of fit, mean standard deviation and so on, and optimized by grid method, and the optimal random forest model is obtained. On this basis, the optimal random forest algorithm is used to predict the number of tourists in the scenic spot in the future. On the first day, the actual tourist flow is 919 and the predicted tourist flow is 924. In this paper, the forecast model of tourist traffic data based on random forest algorithm is convenient for the corresponding management of scenic spots.