高性能无人机平台凭借灵活的作战空域、丰富的信息资源、强大的计算能力等优势逐渐成为了空战中的新质力量，本文针对远距协同攻击非合作目标面临的机载任务决策问题，提出了一种基于轨迹数值优化的机-弹协同导引算法。首先，建立了机-弹远距协同作战的质心运动微分动力学方程，并设计了轻量化数值积分算法快速计算机-弹的轨迹序列。其次，设计了时间-空间-角度约束下导弹轨迹的高效二次型数值优化求解算法，并给出了数值形式下的矩阵迭代运算格式。然后，采用改进Kmeans聚类算法将预测命中区凸包划分为导弹协同攻击子区域，并根据动态观测量对非合作目标的命中子区域进行概率计算与更新。最后，考虑了区域协同攻击、波次同点攻击两种协同方式，设计了机-弹最优协同导引的目标函数及约束方程，来优化机-弹协同作战能力。仿真实验验证了所提算法的正确性及有效性，高性能无人机完成了对非合作目标的区域协同与波次协同的导引决策且脱靶量不大于1 m，所提数值优化算法在嵌入式环境下耗时不大于150 ms，具有一定的工程应用参考价值。

High performance unmanned aerial vehicle platforms have gradually emerged as a formidable asset in air combat， owing to their flexible combat airspace， abundant information resources， and powerful computing capabilities. This paper proposes an aircraft-missile collaborative guidance algorithm based on trajectory numerical optimization to address the airborne task decision-making problem faced by long-range collaborative attacks on non-cooperative targets. Firstly， the differential dynamic equations of the center of mass motion for long-range coordinated combat between aircraft and missiles were established， and a lightweight numerical integration algorithm was designed to quickly compute the trajectory sequence of the aircraft-missile. Secondly， we design an efficient quadratic numerical optimization algorithm tailored for missile trajectories under time-space-angle constraints， alongside providing a numerical format for matrix iteration operations. Then， the improved K-means clustering algorithm is used to divide the predicted hit area convex hull into missile coordinated attack sub regions， and probability calculations and updates are performed on the sub regions of non-cooperative targets based on dynamic observations. Finally， two collaborative methods， regional coordinated attack and wave same point attack， were considered， and the objective function and constraint equation for the optimal coordinated guidance of the aircraft missile were designed to optimize the collaborative combat capability of the aircraft missile. Simulation experiments have verified the correctness and effectiveness of the proposed algorithm. HPUAVs have completed guidance decisions for regional and wave coordination of non-cooperative targets with a miss distance of no more than 1 m. The proposed numerical optimization algorithm takes no more than 150 ms in embedded environments， and has certain engineering application reference value.