Aiming at the low efficiency of existing mobile robot path planning methods, a path planning method based on improved fuzzy adaptive genetic algorithm was proposed. Based on domain knowledge, the feasibility of the initial path was screened to improve the proportion of feasible paths. Fuzzy logic controller was used to dynamically adjust the running parameters of genetic algorithm to improve the speed of path optimization and avoid falling into the local optimal path; the operation safety requirements of the robot was considered comprehensively, the cosine function smoothness evaluation factor was introduced, and different penalties were applied to different path angles to improve path smoothness. The simulation results show that the improved algorithm is effective to solve the path planning problem.