Unmanned Aerial Vehicle (UAV) swarm systems offer unique advantages in applications like disaster relief and wide-area monitoring, leveraging flexible deployment and swarm intelligence. However, traditional scheduling faces significant challenges: the challenge of optimizing task allocation while respecting the swarm’s current positional constraints, the lack of UAV-specific resource models (e.g., battery, position), and inadequate multi-stage task coordination. To address these challenges, we propose a novel UAV swarm scheduling algorithm based on a K3s cluster architecture. The algorithm integrates three core modules: 1) A dynamic node evaluation module assessing capabilities using real-time UAV status (battery, position, latency, hardware) and task requirements. 2) A hybrid multi-objective optimization module employing a greedy strategy for initial solutions, followed by NSGA-II to optimize trade-offs among coverage area, energy consumption, and latency. 3) A task collaboration module managing multi-stage coordination, utilizing UAV locations and scores to select support nodes and deciding task execution location (local vs. nearby) based on task and UAV situation. The resulting framework jointly optimizes otherwise conflicting objectives. Experimental results demonstrate that the proposed algorithm significantly outperforms baseline methods in key metrics like coverage efficiency, energy consumption, and task completion latency, providing an effective solution for UAV swarm operations.