In order to improve prediction accuracy of ship traffic flow,an improved model was developed to predict ship traffic flow based on linear growth model in consideration of all periodic fluctuation factors,such as season,climate,and so on,then the Bayesian estimation and prediction were used to solve the new model,and ship traffic flow was predicted by using the time series data of ship traffic flow. Results show that the proposed model more accords with the actual situation of traffic flow comparing with the linear growth model,and the mean absolute error of monthly ship flow decreases 3.56%,and the standard deviation decreases 3.79%,therefore,it is effective to predict ship traffic flow.