Optimizing land use structure is currently a hot research topic in the land resource management field. In this study, we constructed a new land use structure optimization model based on the improved multi-objective differential evolution algorithm. This model made improvements in two aspects of the classic differential evolution algorithm, i.e., control parameters and adaptive strategies. On this basis, we established a multi-objective function by taking ecological benefits and economic benefits as the objectives. According to the real situation of the study area, we established multiple constraint conditions and finally established an improved multi-objective differential evolution model. By taking the year 2010 as the base period, we simulated an optimized land use quantitative structure in 2020 for the study area and compared this optimized structure with classic linear programming. The experimental results showed that although the annual ecological benefits in the study area decreased by 5,105,300 yuan, the annual economic benefits increased by 69,133,500 yuan, and the annual total benefits increased by 20,878,300 yuan – an increase of 0.44%. This showed that the land use structure obtained by using the optimization model proposed in this paper was more reasonable. The results indicated that the model established in this study possessed quite good properties and could meet the requirements for the regional land use structure optimization under multiple constraint conditions. The optimized results can provide the scientific basis for formulating appropriate measures for regional land resources use.