As an effective approach to penetrate air defense systems, fractionated cooperative attack relies fundamentally on optimizing multi-projectile collaborative trajectory planning capabilities. This study addresses the multidimensional complex constraints in coordinated strikes by Distributed Glide Bomb (DGB) clusters, including initial condition disparities resulting from post-separation dynamics (altitude differences, velocity deviations, and attitude dispersion) and three-dimensional dynamic no-fly zone avoidance requirements. A novel trajectory planning methodology is proposed, integrating the Gauss Pseudo spectral Method (GPM) with coordinated-variable control. By establishing a three-degree-of-freedom (3DOF) kinematic model that deeply couples nonlinear aerodynamic parameter variations with multi-constraint conditions (including dynamic equations and geometric constraints), a multi-objective trajectory optimization framework is developed for DGB cluster coordination. The method innovatively employs an adaptive node distribution strategy to optimize the discretization process of the Gauss Pseudo spectral Method (GPM). By implementing mesh refinement in no-fly zone boundary regions, it effectively resolves the issue of high constraint violation rates at trajectory inflection points observed in conventional approaches. Numerical simulations demonstrate that under heterogeneous initial conditions (varying altitudes, velocities, and attitudes), all DGBs successfully generate collision-free 3D trajectories avoiding complex no-fly zones while achieving synchronized strikes with terminal velocities within specified tolerances. The results validate the method’s effectiveness in multi-constraint resolution and its superiority over conventional approaches in computational efficiency and mission adaptability.