Passengers book their China high-speed railway tickets in advance, and a high-speed railway ticket confirms which seat will be booked. To improve the service quality, the present high-speed railway booking system allows passengers to book preferred seats if there are sufficient tickets. However, this booking strategy results in many isolated seats which are not preferred by the early-bird passengers and cannot meet the need of the later-coming travel groups. To solve this unbalance ticketing selling, this study proposes a new approach which designs a seat assignment to match the supply with the demand to utilize the train capacity with consideration of the passenger seat preference and stochastic booking process. The proposed approach is compared with the genetic algorithm which is a popular heuristic to solve complex problems, and the results show that the performance of the proposed approach is better.