In order to predict the traffic flow of the city comprehensively, objectively and accurately, the monthly traffic flow data of Sanya from 2012 to 2017 were collected for analysis.On the basis of data collation, ARIMA model and gray Markov model are mainly adopted to carry out fitting simulation for the traffic flow of Sanya from January,2012 to December,2017, and to extrapolate the passenger flow trend of each month in 2018.The results show that the average absolute percentage error of ARIMA prediction model is 4.42%, and the average absolute percentage error of Markov model is 3.78%,which indicate that the prediltion by the two models has high precision.Finally, the trend extrapolation is predicted by using the gray Markov model, and the traffic flow in Sanya in 2018 is estimated to be nearly 36 million.The predicted results can play a positive role in the formulation of policies in the tourism, transportation and other industries in Sanya.