This paper establishes a Bayesian network model to investigate the influence of land use characteristics on the travel mode choice behaviors of residents. Travel data and Land use data for this study are collected in Nanjing central city, China. The Greedy Thick Thinning algorithm combined with Receiver Operating Characteristic(ROC) curves are employed for Bayesian network structure learning while the Bayesian estimation method was proposed for parameters learning, and K-fold cross-validation method is implemented to test the model from the d...