Order-type strike plays a crucial role in attacking time-sensitive targets during the Russia-Ukraine conflict. This novel warfare tactic imposes stringent requirements within short windows of superiority, which traditional pre-planning and execution schemes cannot meet. This paper focuses on attacking time-sensitive targets with shorter striking time, lower weapon costs, and higher striking rewards. An improved NSGA-II algorithm is proposed, which includes initializing populations with matching relationships, employing adaptive crossover and mutation operators, eliminating duplicate individuals when combining parent and offspring populations, and integrating particle swarm optimization to improve optimization accuracy. AHP-TOPSIS is applied to evaluate Pareto solutions and select the optimal scheme. Experimental results demonstrate that the improved NSGA-II can reduce the runtime by 27% on average. Compared to the original NSGA-II, the optimal solution of the improved NSGA-II for a specific scenario demonstrates a 9% reduction in strike time, an 8% decrease in cost, and a 3% increase in reward.