Smart card systems have been applied to fare payment of public transport systems. When individual railway passengers who hold a smart card of public transport passes through a ticket gate at a station, the accurate time of his/her passage is recorded in the smart card system. The recorded data of smart card system is stored in a railway company and it becomes a huge amount of historical data. If the time of passing through ticket gates at both origin and destination stations is combined with train schedules, it becomes possible to identify trains boarded by each passenger. The aim of this paper is to develop a methodology to estimate trains boarded by each passenger who is recorded on the smart card system. This paper shows the methodology for the estimation and an algorithm for its computation.