［Purposes ］ To address the issues of short flight endurance，low payload，and ground traffic limitations in single unmanned aerial vehicle （UAV），a two -level collaborative logistics distribution structure using heterogeneous UAV swarms across the entire airspace was proposed.This aims to enhance delivery efficiency through hierarchical resource scheduling，thus breaking the limitation in the traditional “UAV + vehicle ” collaboration mode.［Methods］ A two -level collaborative scheduling model was established with the goal of minimizing the total flight distance.Large transport unmanned aerial vehicles （l-UAVs） were responsible for long -distance transfers and the launch/retrieval of express unmanned aerial vehicles （e-UAVs），while e -UAVs handled the final delivery.A three -phase hybrid heuristic algorithm was designed：a dynamic weight constraint K-means++ clustering algorithm was employed to divide customer points into clusters and determine launch points；an improved ant colony optimization algorithm （ACO） was used to optimize the paths of l -UAVs，and a tabu search algorithm （TS） was adopted to plan the delivery paths of e -UAVs.［Findings］ Solomon dataset -based experiments show that the ACO + CS combination algorithm takes the lead in flight distance and time optimization.Sensitivity analysis indicates that the system performs best when e -UAVs have a payload of 8 kg，proving the model ’s robustness.［Conclusion ］ The proposed “l-UAV + e -UAV” two -level collaboration mode avoids ground traffic bottlenecks via the entire airspace link.Its dynamic recovery mechanism and combined optimization algorithms significantly shorten delivery time，offering an efficient scheduling scheme for low -altitude logistics networks.