Abstract: Aiming at the problems existing in the Changqi Irrigation District of Guangxi, such as aging projects, imperfect irrigation-drainage systems, low informatization level, and insufficient water-allocation accuracy caused by experiencedependent regulation, this paper develops a coupled water-resources allocation model system integrating runoff forecasting,optimal water-resources allocation and precise sluice regulation for Changqi Irrigation District, in combination with the requirements of digital-twin irrigation-district construction and refined water-resources management. Firstly, the three-source Xin'anjiang model is adopted to establish the inflow runoff forecasting model, and the SCE-UA global optimization algorithm is used for parameter calibration and validation. Secondly, an optimal water-resources allocation model is built, which incorporates reservoir regulation, recursive water allocation at canal network nodes, and priority-based supply-demand matching. The model integrates reservoir water balance, canal water attenuation, differentiated water supply and reverse water replenishment mechanisms. Finally, supporting canal-system hydraulic calculation and differentiated sluice regulation models are constructed to convert water-allocation schemes into actual engineering operating conditions. Taking the peak spring-irrigation period in early April 2025 as a case study, application research is carried out. The results show that the Nash-Sutcliffe efficiency coefficient of the runoff forecasting model reaches 0.73 in both calibration and validation periods, and the simulation accuracy meets practical operational requirements. The water-resources allocation model realizes multi-source joint regulation and refined node-by-node water allocation; the water supply process of main canals remains stable, and the spatiotemporal characteristics of water distribution in branch canals are highly consistent with actual water demands. The sluice regulation model can output daily opening degrees and operation durations for each sluice gate, and its regulation logic conforms to engineering operation rules. The research outcomes can provide a case demonstration for digital-twin construction and intelligent water-resources regulation of irrigation districts.