Owing to the high-dimensional non-convexity and strong coupling among UAVs, existing approaches to this problem are prone to iterative oscillations, making it difficult to accurately obtain the optimal trajectories. To address this issue, this paper proposes a cooperative trajectory planning method that integrates time-urgency asymmetric decoupling with adaptive sequential convex programming. A bi-level planning architecture that combine the cooperative decouple and solve trajectory is constructed. At the cooperative level, an asymmetrical decoupling strategy based on time urgency is adopted to dynamically generate priority sequences, thereby decomposing the high-dimensional strong coupled UAV trajectory planning problem into several single-UAV convex programming subproblems. At the solution level, a virtual corridor is introduced to sparsify collision constraints and reconstruct the effective constraint set within these convex programming subproblems. The signed distance field is employed to convexify the inter-UAV collision and threat constraints. To mitigate the convex approximation error, the terminal and collision constraints are transformed into penalty terms in the objective function. Meanwhile, an adaptive trust-region adjusts the deviation between the reference and optimized trajectories, which enables the achievement of optimal cooperative trajectory planning for multiple UAVs. Simulation results demonstrate that the proposed algorithm improves trajectory planning quality and efficiency compared to the conventional SCP and improved SCP methods.