农业灌区是农业最主要的用水大户，其节水水平直接关系到整个农业节水目标的实现。本文选择河北省石津灌区为研究区域，根据灌区的实际情况，提出基于云模型和可变模糊聚类迭代模型的模糊综合评价法，有效地克服了隶属度确定的随机性和模糊性，同时权重确定也更加合理。利用该方法评估了灌区2002年~2011年的节水水平，评估结果表明，灌区节水水平逐年提高，由耗水型转变为过渡型，但灌区渠系水利用系数和灌溉水利用系数整体偏低，有待于进一步提高。研究成果可为大型灌区节水评价以及灌区的节水管理提供技术支持。

Agricultural irrigation area is the main water user of agriculture, and its water-saving level is directly related to the realization of the whole agricultural water-saving goal. Shijin Irrigation District of Hebei Province was chosen as the study area and according to its actual situation, the fuzzy comprehensive evaluation method based on cloud model and variable fuzzy clustering iterative model was proposed to assess the level of water saving in the irrigation district from 2002-2011. This method overcame the randomness and fuzziness of the membership degree determination effectively, and made the weight determination more reasonable as well. The results showed that the water saving level of the irrigation area improved year by year, and the water saving type changed from consumption to transition. However, the whole utilization coefficients of canal-system water and irrigation water were low and needed to be further improved. The research results can provide technical support for water saving evaluation of large-scale irrigation districts and water-saving management of irrigation districts.