To address the real-time and efficiency requirements for multi-target cooperative interception by unmanned combat systems in dynamic battlefield scenarios， this paper proposes a greedy target assignment and fire scheduling method based on dynamic time window （GTAFS-DTW）. By constructing a relative motion model between combat units and targets， the proposed method incorporates engagement distance constraints， task execution constraints， and target maneuver characteristics. a dynamic time window screening mechanism is designed with the optimization objective of minimizing engagement time and impact time difference. On this basis， a hierarchical greedy strategy is adopted to prioritize unit-target pairs with overlapping feasible engagement intervals， and temporal cooperative optimization is achieved through iterative adjustment of window boundaries and fire scheduling. Simulation results demonstrate that GTAFS-DTW effectively reduces impact time deviation to within 0.5 s in both lateral and longitudinal queue scenarios compared with traditional discrete particle swarm optimization （DPSO）， significantly enhancing computational efficiency and satisfying real-time optimization requirements.