Abstract: For the public bicycle system vehicle scheduling problem, this paper solves the problem by analyzing public bicycle demand characteristics at different times of a day. Based on the objective of minimized costs and maximized satisfaction, two models are created, i.e., public bicycle vehicle allocation path optimization model without time window and public bicycle vehicle allocation path optimization model based on rolling horizon, and ant colony optimization is used to solve the model. Finally, the example result shows that a reduction of 48.5% and the decrease of 14.1% at flat peak are attained with the optimized route, while an increase of 62.3% at peak period is achieved with the optimized satisfaction. It shows that the models are reasonable and scientific.