Based on the problem that the traditional road information detection method cannot obtain real-time continuous road information, a method of using the unmanned aerial vehicle (UAV) to conduct road network inspection was proposed. The optimization model for multi-machine flight path was established to solve the path optimization problem of the UAV through the space-time road network. The optimization objective for the new model was to minimize the total flight time of all aircraft or the maximum flight time of each aircraft under the condition of the completion of all mission. The new model not only transformed the dynamic path plan into a static path plan by using time-space network technology to meticulously depict the flight path of the UAV during the patrol process, but also added multiple patrols of key road sections and time interval constraints for multiple patrols. Based on the analysis of a case in the text, results show that, in comparison with the path planning without considering the number of patrols, the total and single flight time of the UAV increases by 15.87% and 15.15% respectively. The mission goal of patrolling two important road sections for three times can be completed. The analysis of the example shows that the optimized patrol path is more realistic to practical needs.