With the accelerated process of urbanization and traffic development, especially the urban rail transit system's great improvement, Park-and-Ride provides an effective mode for trips between suburbs and downtown. In this research, online and field survey is carried out on the use of Park-and-Ride facilities. Analyses are conducted on personal attributes containing gender, age, and income; the travel characteristics such as driving time during departing from origin to parking lot, parking duration, transfer mode, transfer walking time and waiting time, and transfer times; Park-and-Ride users' intentions concerned walking time, waiting time, and time looking for parking space; and reasons for Park-and-Ride trip mode not be chosen. On the basis of decomposition for travel procedures, impedance models for different trip modes including public transport, private car, and Park-and-Ride are built and then the multinomial logit model for choice probability of trip modes and Park-and-Ride demand model is established. After further analysis on the survey data, calibrations and tests for the impedance models above are performed. Finally, a case is shown to demonstrate application of the proposed model.