In order to ensure the riding safety of users and the normal operation and maintenance of enterprises, it is necessary to recover the unusable sharing bicycles in time, and scientifically plan the recovery path to improve the recovery efficiency. Based on a brief analysis of the current situation of recycling, the unusable sharing bicycles are defined and classified, and the recycling criteria and specific procedures are determined. Next, the K-means algorithm is chosen to cluster the unusable sharing bicycles. A path optimization model with the lowest total cost of recovery work as the goal was constructed, and an improved ant colony algorithm was designed to solve the problem by taking the faulty Mobike bicycles in some areas of Xuhui District, Shanghai as an example, and the correctness and effectiveness of the models and algorithms used were verified. Research has shown that the full load factor is positively correlated with the travel time of recycled vehicles, and the number of service nodes affects both the transport time and travel time, which has a significant impact on the total cost of recycling. Therefore, when arranging recovery tasks, companies should take into account the real situation of the workload and planning costs to select specific parameters, so that the model more closely matches the actual planning needs.