In recent years, as drought intensifies and agricultural water consumption increases, it is of great significance to optimize the irrigation schedule to ensure regional food security. This paper constructs the distributed AquaCrop model and multi-objective genetic algorithm (NSGA - Ⅱ) simulation - optimization (MGSO) model to facilitate the development of a rational irrigation schedule. The distributed AquaCrop model considers the spatial variability of soil, climate, crops, and management practices, which can be batch calibrated using the XGBoost method. The MGSO model is for irrigation schedules under the combination of different Precipitation and ET₀. In this paper, crop yield, ET, and water use efficiency (WUE) were simulated and analyzed in 13 irrigation zones in Northeast China, where the existing irrigation schedules were analyzed and optimized. The results showed that the distributed AquaCrop model could simulate regional crops well. Crop yields in the study area ranged from 3 to 10ton/hm². The western part of Heilongjiang province and the northern part of Jilin province has a higher yield. The simulation results of Heilongjiang Province are more accurate, and the relative error is minor. The joint distribution model constructed by the Frank Copula function can describe the joint probability distribution characteristics of precipitation and ET₀. According to the simulation results, each typical station has a different performance under the existing irrigation schedule under different situations. The crop yield and WUE of some stations changed significantly. The maximum and minimum yield difference was 22% for Harbin, 32% for Heihe, and 21% for Dunhua. It is mainly due to the irrigation amount in some scenarios that do not meet crop water requirements. Under the optimized irrigation schedule, the crop yield and WUE in different scenarios have been improved by the MGSO model in the Harbin station

近年来，随着干旱加剧，农业用水量增加，优化灌溉计划对保障区域粮食安全具有重要意义。本文构建了分布式AquaCrop模型和多目标遗传算法（NSGA-Ⅱ）模拟-优化（MGSO）模型，以促进合理灌溉计划的制定。分布式 AquaCrop 模型考虑了土壤、气候、作物和管理实践的空间变异性，可以使用 XGBoost 方法进行批量校准。MGSO 模型是针对不同降水和 ET ₀组合下的灌溉计划. 本文对东北13个灌区的作物产量、ET和水分利用效率（WUE）进行了模拟分析，对现有的灌溉计划进行了分析和优化。结果表明，分布式AquaCrop模型能够很好地模拟区域作物。研究区域的作物产量在 3 到 10 吨/小时^{2 之间}。黑龙江省西部和吉林省北部产量较高。黑龙江省的模拟结果更准确，相对误差较小。Frank Copula 函数构建的联合分布模型可以描述降水与ET ₀的联合概率分布特征. 根据模拟结果，每个典型站点在不同情况下在现有灌溉计划下具有不同的性能。部分站点作物产量和WUE变化较大。最大和最小产量差异为哈尔滨22%，黑河32%，敦化21%。主要是由于某些场景下的灌溉量不满足作物需水量。在优化的灌溉计划下，哈尔滨站MGSO模型提高了不同情景下的作物产量和WUE