Abstract: To address the relatively extensive management and low allocation efficiency of water resources in irrigation districts of southern China’s hilly regions, this study focused on a typical irrigation district in Suxian District, Chenzhou City, Hunan Province, and developed an intelligent water resource allocation and decision-support platform for refined irrigation management. The platform applies multi-source spatiotemporal fusion methods, including Flexible Spatiotemporal Data Fusion (FSDAF) and the Time Series Linear Fitting Model (TSLFM), to integrate Landsat and MODIS data, generating high spatiotemporal resolution vegetation index series that drive the Priestley–Taylor Jet Propulsion Laboratory (PT-JPL) model for dynamic simulation and prediction of evapotranspiration. A multi-objective optimization model using the Non-dominated Sorting Genetic Algorithm Ⅱ (NSGA-Ⅱ) was constructed to minimize both irrigation water consumption and irrigation deficit rate. The entropy weight method combined with the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) was applied to screen and select optimal irrigation scheduling schemes. Furthermore, by integrating the MATLAB computational engine with front-end visualization technologies such as Vue.js and OpenLayers, a smart decision-support system was developed, integrating evapotranspiration simulation, irrigation demand analysis, and water allocation optimization. Results indicate that the system can achieve dynamic evapotranspiration simulation and optimized water allocation, effectively improving the scientific rigor and refinement of irrigation scheduling. The system provides technical support for water-saving management and efficient water use in irrigation districts of southern hilly regions and has potential for practical application and wider promotion.