Abstract: During peak holiday periods when passengers are concentrated, the act of buying short distance ticket for long-distance travel has become a difficult problem for passenger transport organization and management. This paper proposed a prediction model for passenger's act of buying short distance ticket for long-distance travel based on identifying popular destinations during peak period. The paper analyzed historical passenger flow patterns and urban travel popularity, implemented the risk probability assessment of short distance passengers buying short distance ticket for long-distance travel on popular train numbers, selected actual train operation and ticket replenishment data during peak hours in 2023 to test the model. The results show that the overall mean square error was 0.26%. It indicates that the model has practical application conditions. The trial results show that this model can provide effective decision-making basis for passenger transport management departments to ensure the safe operation of trains during peak hours.