In order to improve the efficiency of mobile robot autonomous inspection in human-machine mixed environment, we proposed an obstacle avoidance algorithm based on fuzzy logic theory. By analyzing the rules of human social behavior and combining the motion characteristics of the robot, established a mathematical model of mobile robot system. The human-machine distance, human behavior patterns, and relative speed parameters between human beings and robots are fuzzy so as to get the decision of robot speed regulation based on fuzzy inference method. Finally, the simulations and experiments of two motion behaviors of crossing and encountering in the man-machine hybrid environment are discussed and analyzed, and the results verify that the algorithm has anti-interference and good obstacle avoidance performance.