Abstract: This paper addresses the trajectory transfer and proximity flight problems of multi-spacecraft swarms operating in large-scale, complex, and unstructured space-station environments. A cooperative trajectory planning method is proposed, combining long-range minimum fuel transfer with attitude-constrained safe flight in proximal regions. Based on Fibonacci-sphere sampling and the Clohessy–Wiltshire (CW) equations, a fuel-optimal cooperative transfer is achieved from the initial states of multiple spacecraft to the circumscribed sphere of the space station. An A*-based path search algorithm with attitude-feasibility checking is employed to identify narrow regions that require specific attitudes for traversal, based on which a safe flight corridor (SFC) is constructed. Taking the SFC as a hard constraint and the CW equations as soft constraints, dynamically feasible trajectories are generated using a minimum-snap optimization method. Simulation results demonstrate that, compared with conventional spherical-envelope planning and fixed-time transfer strategies, the proposed method effectively reduces fuel consumption and significantly improves spacecraft navigability through narrow structural gaps as well as trajectory smoothness. The method is suitable for cooperative proximity-flight tasks of multi-agent systems in unstructured environments and provides reliable technical support for future space-station on-orbit servicing and cluster operations.