Abstract: The influencing factors of rail transit passenger flow are a focus of rail transit research. The spatio-temporal nonstationarity of passenger flow at different stations was considered to be related to the built environment of station area. A spatio-temporal geographically and temporally weighted regression (GTWR) model was constructed to reveal the impacts of land diversity, density and site attributes on Tianjin rail transit passenger flow in time and space dimensions. Results show that compared with the traditional geographically weighted regression (GWR) model and ordinary least squares (OLS) model, the GTWR model has better goodness of fit. The density of bus stations promotes the passenger flow of rail transit, especially in the morning and evening rush hour of working day and the location of central city. Business facilities in the city centre attract more subway passengers at workday evening peaks, while in the suburbs they attract more subway passengers at morning peaks. Population density promotes passenger flow in rail transit. Adequate parking facilities can attract more rail passengers.