Abstract: To improve the safety and timeliness of cooperative motion of multiple intelligent firefighting robot systems in complex fire environments, a formation obstacle-avoidance control method is proposed, which combines the virtual leader-leader-follower method and an improved artificial potential field method using a dynamic weight. Firstly, a cooperative formation controller combining a circular motion control law and a bearing angle based positioning control law is designed, to make each robot converge to its desired position along a circle with the virtual leader as the center. Then, the traditional artificial potential method is improved, and a logarithmic obstacle function is used to establish the dangerous expansion areas on both sides of the road to ensure that the formation drive in a safe area. The direction of the resultant repulsive force from the obstacle and the road boundary is adjusted to be perpendicular to the direction of the attractive force from the target and far away from the obstacle, so as to solve the problems of local minimum and unreachable target. Finally, a dynamic weight factor is introduced to adaptively adjust the ratio of the formation controller to the obstacle avoidance controller. To validate the effectiveness of the proposed controller, it is compared with the traditional artificial potential field method and an improved artificial potential field method with fixed weight by simulation experiments. The results show that the designed controller outperforms the other methods regarding convergence speed, tracking error, and obstacle avoidance effect. Further, 3 intelligent firefighting robots are used in a physical experiment to validate the practicality of the controller. The result shows that the controller can effectively restrict the formation within the feasible road area and complete the obstacle avoidance.