Based on the background of manned robots transporting passengers in large public indoor environments such as terminal buildings, this paper proposes a model predictive obstacle avoidance control method based on control barrier function to autonomous manned mobile robots’ obstacle avoidance in pedestrian dense environments and improve their engineering security. Firstly, the safety constraint was turned into the set's forward invariance by controlling the barrier function to achieve the safe obstacle avoidance distance. Secondly, to account for the unpredictability of the obstacle volume, the minimum distance was introduced using the control barrier function to ensure further safety. In order to achieve safe obstacle avoidance by manned robot. Simulation and experimental results demonstrate that the proposed method is able to circumvent obstacles in real time, and the resulting obstacle avoidance trajectory displays smoothness.