This paper addresses cluster collaborative trajectory planning and proposes a spatiotemporal multi-constraint cluster trajectory optimization method based on a space-based optimization strategy. The method involves the design of a task planning layer and a trajectory planning layer. The task planning layer employs an improved hybrid particle swarm algorithm for terminal point planning, while the trajectory planning layer, building on the Dubins method, introduces radius optimization strategies and trajectory reconfiguration and optimization strategies based on virtual terminal points. Through these two-layer optimization methods, the cluster can simultaneously reach the terminal point, achieving collaborative trajectory planning under multiple spatiotemporal constraints. Additionally, this approach effectively reduces computational complexity and enhances the engineering practical value of the algorithm. Simulation verification demonstrates that the algorithm proposed in this paper exhibits excellent adaptability and effectiveness under random initial conditions. Compared to the traditional improved Dubins method, its adaptability is significantly enhanced, enabling trajectory variance to converge to the ideal range under any initial conditions. Furthermore, it offers advantages such as computational simplicity and good real-time performance.