This paper applies neural network modeling approach to analyze the impact of passenger activities on bus dwell time and station-to-station travel time. Data used to develop the model was collected by onboard AVL/APC devices. Sensitivity analyses based on a trained neural network were performed to evaluate the relative significance of each passenger activity variable to variation of dwell time and/or station-to-station travel time. Transit providers can use these methods to identify the causes of schedule deviation and to develop improvement measures that are most effective to transit service.