For multi-unmanned surface vehicle (USV) systems navigating through narrow channels and other complex environments, a formation obstacle-avoidance control algorithm is presented that integrates the artificial potential field (APF) method with model predictive control (MPC). To mitigate these problems，the traditional APF approach is improved, including the use of a saturated gravitational potential field and a partitioned repulsive potential field. These modifications aim to improve the precision of obstacle avoidance and the ability to maintain formation coherence in complex environments. Furthermore, by leveraging the multi-step predictive optimization capabilities of MPC, the proposed algorithm dynamically adjusts control inputs based on desired trajectories generated from potential field forces, which ensures the stability of formation control and the effectiveness of obstacle avoidance, thereby avoiding the path oscillation issues encountered by traditional artificial potential field methods in narrow waters. Simulation results demonstrate that the proposed algorithm outperforms traditional methods in terms of obstacle-avoidance success rate, formation stability, and path planning efficiency. And the improved algorithm can avoid local minima, maintains formation integrity, and ensure smoothly passage through narrow channels, thereby validating its effectiveness.