Nowadays, water scarcity is getting increasingly severe, particularly in arid and semiarid areas of China. In addition, there are multiple uncertainties and nonlinear issues in the water resources management process. To address these issues, we developed an improved fuzzy multi-objective nonlinear model incorporating Particle Swarm Optimization and fuzzy programming. The model was applied to a case study in Minqin County, Gansu Province, China, with the objectives of maximizing economic benefit and water productivity. Based on the developed model, a series of irrigation water resources optimal schemes under different alpha-cut levels were provided. From the results, as the alpha-cut levels increases from 0 to 1, the optimal results for economic benefits, water productivity, and crop irrigation areas show a consistent trend of decreasing upper bounds, increasing lower bounds, and narrowing intervals until a deterministic value occurs at alpha = 1. Furthermore, the optimal results of the developed multi-objectives model are always in the middle of the optimal results of the individual objectives, which represents that the established model has the ability to deal with multiple conflicting objectives and provide decision-makers with a rang of optimal schemes under uncertainty.