Unreasonable site location selection of intercity railway stations usually results in high investment and poor attraction of passenger volume. For that the previous studies neglected or simplified the calculation of broadsense travel utility in a city (a part of tripchain utility), the paper proposes an innovative model on the basis of the tripchain characteristics and employs the logit model to estimate the probabilities of various travel modes along the trip chain. A bi-level objective function based on the trip chain is developed to maximize the incomes of intercity rail operations (the upper level) and minimize the costs of passenger travels (the lower level). The improved particle swarm algorithm is utilized to find the optimum solutions. The results of the case study on the railway site location selections in Zhuhai City indicate that the Pareto optimum of model suggests the contradictive relationship between the objective functions at the upper and the lower levels, which is essential to identify the optimum tradeoff among the various demands of the rail stations. The model serves to provide a reasonable decisionmaking procedure for the planning of the railway site location selections in the future.