Abstract: In order to avoid the obstacle collision of monorail transport vehicles in mountain orchard and improve the safety and stability of orchard operation, a set of obstacle avoidance system of monorail transport vehicles in mountain orchard was designed by using depth camera based on crossing time method. When the system is running, the depth camera obtains the information of obstacles in the passage of the transport vehicle, and processes the data through the Raspberry Pi on-board computer running the Linux system. After making a decision, it sends a signal to the transport vehicle control center Stm32 microcontroller to control the motor change Movement status. The experimental results show that the system has an obstacle recognition rate of 100%, and for interference-type side branches that meet the pruning requirements, the false touch rate is below 8%, the minimum braking distance for obstacle avoidance is 101 cm, and the maximum delay of the system is 0.475 s.