During the war, equipment is constantly being damaged with limited battlefield rush-repair time and power. Therefore, some military problems are presented in this paper. In order to get more fighting time for damaged equipment to participate in operation again as much as possible, three problems should be considered properly. The first problem is how to dynamically choose the most suitable damaged equipment for each repair group. The second one is how to divide tasks between different groups. The third one is how to determine execution sequence in the same group. A mathematical model is established to solve the dynamic battlefield rush- repair task scheduling problem （DBRTSP） in wartime. A variant genetic algorithm is designed to dynamically track the change of the optimal solution. A scheduling example is solved through Matlab. Results show that the proposed model is not only scientific and reasonable, but also convenient and efficient.