To meet complex space-time coordination constraints, it is crucial to adjust the formation configuration when the wave gliders (WG) formation performs cooperative tasks in challenging marine environments. The propulsion of the WG relying on wave energy leads to nonlinear and unpredictable motion control, to address this challenge, a flexible formation path planning method based on the constraint-segmentation and multi-objective particle swarm optimization (CS-MOPSO) algorithm was proposed. A divide and conquer strategy was used to segment a pre-planned path into multiple waypoints, and the formation configuration of each waypoint were optimized by the CS-MOPSO. The CS-MOPSO algorithm segments the particle population according to constraint violation level and guides particles toward feasible region. Dubins curves were employed to construct collaborative paths with flexible formation constraint. A formation configuration evaluation system was proposed for quantitative assessment of the formation configuration and cooperative path, and the technique for order preference by similarity to the ideal solution was used to identify the optimal solution. Simulation results show that compared to other algorithms, the proposed method achieves shorter navigation time and lower formation distortion under complex constraint.