As one of the important technologies of active management and control for highway, ramp metering can effectively alleviate congestion problems in merging areas of highways. A reasonable control strategy is the premise to ensure the control effect. This paper proposes a type-2 fuzzy logic control method optimized by the simulated annealing aided genetic algorithm (SAGA) with the Kriging model based on the infill sampling strategy, taking into account the highway operational performance of main line and ramp. The method uses type-2 fuzzy logic for inference of the control strategy, and combines the variable universe control and the search algorithm to adjust fuzzy control parameters. With a view to solving the expensive optimization problem in simulation optimization, a surrogate model updated by the infill sampling strategy is applied. Taking the highway in Jinan City from Shandong Province as an example, the effectiveness and superiority of the proposed method are verified through the comparative experiments, and different control methods are designed for different traffic scenarios. The results show that the control strategy can productively optimize the traffic operational status of the main line within the constraint of the ramp in congested situation. In non-congested situation, it can further improve the traffic operational performance of the main line while saving computational resource and cost.