Abstract: In order to solve the problem of obstacle avoidance of intelligent weeding machines in complex forest weeding operation environment in southern hilly and mountainous areas, a control system suitable for forest weeding and obstacle avoidance operation was designed first, on the basis of previous research.The forest obstacles are sensed by the combination of the structured light camera LeTMC-520 and the AJ-SR04M ultrasonic sensor.The core of the control system is composed of Jetson Nano as the upper computer and STM32F405RGT6 as the lower computer with a fuzzy controller algorithm, of which membership function and fuzzy control rules improved by the wolf pack algorithm（IGWO） and the simulation verification of the algorithm is carried out using MATLAB.Secondly, the feasibility of the scheme and the reliability of the obstacle avoidance function were verified through the field weeding test in the Camellia forest in the south of China.The results show that compared with traditional fuzzy PID controller, this controller is more accurate, and its obstacle avoidance effect can meet the obstacle avoidance requirements of weeding operations in unstructured forest land in the south, and can provide reliable safety protection for the machine.