The work aims to enhance the operational efficiency of urban public transportation facilities, improve the work efficiency of designers, and meet users' travel needs. User needs were gathered through interviews, which were then transformed into design requirements. The AHP (Analytic Hierarchy Process) was utilized to prioritize design features based on weight. User preference prompt words were extracted and combined with similar prompt words obtained from GPT-4.0 regarding public transportation facilities, resulting in integrated prompt words. Relevant large models were selected, and integrated prompt words were used to generate the design on the SD (Stable Diffusion) platform. Taking the smart bus station as an example, smart public transportation facility plans for different scenarios were generated. After screening by experts, an optimized set of plans was obtained. Users were then invited to score these plans by the Likert scale to determine the best plan for each scenario. The human-computer collaborative design of smart bus station plans can meet users' needs and also improve the work efficiency of designers. Therefore, the design practice of smart bus stations has refined the design process of smart transportation facilities supported by AIGC (Artificial Intelligence Generated Content) technology, providing guidance for the design of other types of public transportation facilities.