The non-cooperative game of unmanned aerial vehicle （UAV） swarm in the complex combat environment is an important research field. In this paper， we investigate the dynamic behaviors of UAV swarm in the scenario of air-ground engagement. We analyze the key factors of combat situation， and use the space state matrix to represent the situation of combat. Based on the dynamic changes of combat situation， the dynamic objective functions with adaptive weights are introduced. We propose a model of UAV swarm engagement under incomplete information， and verify the rationality and effectiveness of the adaptive weights. Our work provides new insights into the dynamic decision of UAV swarm.