Aiming at the path planning and control problem of multi-mobile robot gathering, this paper proposes a path planning strategy based on an improved fast marching square method. Firstly, the segmentation function is applied to improve the velocity map to realize safer and more efficient path planning, which can map the velocities on the fast marching grid map to the real robot velocities and reduce the redundant paths generated by the traditional fast marching square method in the process of backtracking the paths; then, for the three task requirements of minimizing the total energy consumption in the gathering process of the multi-mobile robots, maximizing the space near the gathering point, and minimizing the total energy consumption of the gathering process under the constraints of the gathering formation, different objective functions are analyzed and designed to give the gathering point of multi-mobile robots and the corresponding paths, which demonstrates the validity of this paper's method as well as its applicability in different scenarios. Finally, on the basis of the vehicle dynamics model, trajectory tracking simulation experiments are carried out using the model predictive control with the improved speed on the fast marching grid map as the robot's reference speed, and the results show that the tracking error is reduced, which verifies the effectiveness of the improved speed field of this paper, and it can be applied to the gathering path planning and control of multi-mobile robots in the real environment.