Travel mode identification is crucial for traffic planning and management, as it can help optimize the structure of travel modes and relieve road traffic congestion. The present work proposes a Bayesian Network Learning Framework to identify the travel mode of urban residents leveraging cellular signaling data. Travel behavior attributes, personal attributes, and traffic environment attributes are the three types of explanatory elements considered in this situation. For evaluating the independence between model variables and depicting the causal linkages between these variables, a Bayesian network causality diagram structure learning approach that integrates information theory and probability theory is proposed. The next stage is the proposal of a Bayesian network parameter learning technique based on maximum a posteriori estimate. The model fully utilizes transportation network geospatial data, travel survey data, and cellular signaling data from Kunshan City, achieving high accuracy and robustness in case analysis. The manuscript is concluded by presenting the potential of the model in improving transportation planning and management by providing a highly accurate identification rate of the traffic travel mode.