To minimize the general social cost of road passenger transport, it is necessary for the government to optimize the passenger route plans and balance the benefits between transportation companies and passengers, especially for the cities when there are several coach stations. To provide a method that makes the decision-making more reasonable, a passenger route allocation model based on bi-level programming and a solution based on genetic algorithm are proposed. The upper programming embodies the governmental decision-making behavior of passenger routes, while the lower programming embodies the individual choice behavior of coach stations. The application of the model and its algorithm are illustrated with a practical example. The result shows that the model and the algorithm are feasible and effective.