Abstract: Passenger flow forecast is an important basis for railway network planning, railway line and station design, railway operation and so on. In view of different types of railway passenger flow prediction business needs at multi levels of railway enterprise in different periods of time, the railway passenger flow prediction system is built relying on the railway passenger transport big data platform, which can complete the test, evaluation and standardized application of passenger flow forecast schemes on the same platform. The system can integrate various passenger flow prediction algorithms and models and make full use of historical ticket sales data and consider annual, seasonal, week, time, weather, holidays, major events and other kinds of influence factors to provide passenger flow prediction functions for different business needs with different granularity and in different periods of time. Besides, this system can support two application modes of human-computer interaction prediction and timing batch offline prediction and provide customized passenger flow forecast data for passenger transport business analysts, and provide accurate and reliable decision-making basis for passenger transport business such as intelligent pre-allocation of ticket amount and preparation of train operation scheme.