Abstract: In response to the problem of inaccurate estimation of railway passenger travel demand due to limited by ticket capacity, this paper introduced candidate ticket purchase data to study the railway passenger travel demand mining method. The paper sorted out the relationship between candidate ticket purchasing and travel demand, classified candidate ticket purchasing behavior based on business rules and passenger choices, mined passengers' pre-sale travel needs, and calibrated passengers' actual itineraries based on real name registration data and gate card swiping data, thereby mined the upper and lower limits of passengers' final travel demand. The paper selected the urban OD (Origin - Destination) of Beijing - Nanjing for verification analysis. The results show that compared with the existing candidate demand statistical rules, the final travel demand can maintain a similar pattern to the actual travel demand calculated by the gate, which is more reasonable and provides scientific basic data support for passenger demand forecasting and railway daily marketing analysis.