This paper aims to improve transfer utility between bike-sharing and subway. For this paper, the transfer costs of three combined travel modes were analyzed, including “Bike-sharing + Subway”, “Walking + Subway” and “Bus + Subway”, and a transfer cost function, including time cost and expense cost, was constructed. Cluster analysis was carried out on the origin and destination of bike-sharing in the transfer-influenced area. The sum of squares of errors and the service radius were used to verify the clustering results. Then the number of alternative facilities, the location of alternative facilities and the initial number of shared bikes were preliminarily predicted. Based on the initial scheme, a bi-level programming model of facility layout and configuration in bike-sharing was established, with the goal of improving ride volume and reducing the transfer cost and facility-operation costs to optimize the initial facility-allocation scheme. Then a heuristic algorithm was used to solve the model. Finally, a typical subway station was selected as a case, and the configuration process for facilities in bike-sharing is discussed in detail in the paper. The research results of this paper may provide reference for the planning and optimization adjustments of facilities for bike-sharing.