Abstract：
To improve the traditional travel mode choice model and reflect the features of the historic blocks as well as enhance the prediction model accuracy, this paper consider the influence of three aspects including traffic characteristics, travel characteristics and the trip self property for travel mode choice model firstly. Then, it take the travel data of investigation of inhabitants of four historic blocks including Xi'an, Zhengzhou, Kaifeng and Luoyang as the examples to analyze model. Finally, the option model of traffic mode of historic blocks based on BP neural network is built. The results show that the proposed model accords well with real values and the model has practical value, so that it will be an effective tool to provide the basis for traffic demand forecasting and traffic planning.
