A wide variety of environmental management problems are solved with a computationally intensive simulation-optimization framework. In this study, the "model pre-emption" strategy is introduced for increasing the efficiency of solving such multi-objective optimization problems. This strategy makes the optimization algorithm avoid the full evaluation of predictably inferior solutions, is applicable to many optimization algorithms, and does not impact the optimization results. Multi-objective pre-emption is used to optimize a new regulation plan for Lake Superior. The new plan is designed to mitigate extreme water levels and increase the total regulation benefits. The rule curve parameters defining the plan are obtained from a multi-objective, multi-scenario optimization problem. Results show that model pre-emption drastically increases the efficiency by up to 75%. The optimized regulation plan outperforms the current plan under the historical scenario. Notably, the optimized plan successfully handles an extremely dry scenario in which the current plan fails to maintain reasonable lake levels.