This study addressed the pressing challenges of uneven spatial distribution of water resources and the acute supply-demand imbalance in Yuyang District, Yulin City. The district represents a typical and representative region of the Loess Plateau in northern Shaanxi. Drawing on a comprehensive dataset covering the period from 2005 to 2020, the research employed a multifaceted analytical framework that integrated the InVEST (Integrated Valuation of Ecosystem Services and Tradeoffs) model, the quota method, and spatial analysis techniques to examine the spatiotemporal evolution of water supply-demand dynamics in the study area.To improve the accuracy and robustness of water resource allocation forecasting, this study introduced the Dwarf Mongoose Optimization Algorithm (DMOA) to enhance the traditional Particle Swarm Optimization (PSO) approach. A hybrid DMOA-PSO-NSGAⅡ model, incorporating the second-generation non-dominated sorting genetic algorithm (NSGAⅡ), was constructed to optimize water allocation across sectors in Yuyang District with projections to 2035. The empirical analysis yielded several key findings: (1) Between 2005 and 2020, Yuyang District experienced a steady increase in water supply, with an average annual growth rate of 6.7%. A clear geographical gradient was observed, with higher supplies concentrated in the southern areas and lower supplies in the north. At the same time, total water consumption showed a continuous upward trend, primarily driven by growing agricultural and industrial demands, which accounted for the majority of annual water use. Although the overall supply-demand balance improved-reducing high-risk areas to only 8% of the district-densely populated urban centers continued to face persistent water scarcity risks. (2) The proposed DMOA-PSO-NSGA-II algorithm outperformed conventional methods in terms of solution diversity and convergence speed, confirming its effectiveness and adaptability for optimizing water allocation in arid and semi-arid regions. (3) Water supply sources were categorized into surface water, groundwater, and unconventional water, while water use was classified into agricultural, industrial, domestic, and ecological water use. Through the optimization model, the water resource allocation scheme for Yuyang District in 2035 was obtained. This plan prioritized the safeguarding of water for living and ecological purposes, strengthened the utilization of unconventional water sources, and proposed comprehensive management strategies for the sustainable utilization and high-quality development of water resources in Yuyang District.This research provided a solid scientific foundation for achieving rational and sustainable water resource allocation in Yuyang District. Furthermore, it offered valuable insights and a replicable reference model that can inform water management strategies in other small-scale arid regions.