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.