Abstract: In order to predict the railway passenger traffic volume more accurately, this paper used the grey correlation method to analyze the influence degree of different factors on the railway passenger traffic volume, determine the main influencing factors, and took them as the prediction indexes to propose a combined predicting model of railway passenger traffic volume based on Verhulst-RBF neural network.Based on the railway passenger traffic volume data of Sichuan Province in recent 14 years, the combined model was tested.The experimental results show that the prediction accuracy of Verhulst-RBF neural network combined model is higher than that of single Verhulst model or single RBF neural network model.