Multi-agent formation refers to several agents completing a specific task through cooperation and coordination according to the expected formation. When performing formation tasks, each agent has its final target position and needs to maintain the desired formation and effectively avoid obstacles in the process of planning to reach the target point. However, a single formation control method has some limitations, such as difficulty in obstacle avoidance in complex environments and rough formation trajectories. In response to the above issues, the main work of this article is as follows: a formation control method combining flexible navigation following method and enhanced conflict search is used, and a safe corridor is introduced as a constraint condition. An MPC controller is used to solve the nonlinear optimization problem. This method overcomes the limitations of traditional methods and can achieve the effects of formation and obstacle avoidance in complex environments. By using formations, relatively smooth trajectories can be obtained. Taking 6 agents as an example, when using a triangular formation, the overall trajectory optimization rate reaches 9.2412%.