The abstract should summarize the contents of the paper in short terms, i.e. 150–250 words. The purpose of this paper is to discuss the application of clustering algorithm to classify and analyze the agricultural irrigation load curve, so as to better understand the influence of agricultural irrigation demand on the distribution network. The temporal distribution of agricultural water demand is the key to efficient management and optimal scheduling of distribution network resources. Due to the significant seasonality of agricultural irrigation demand and the diversity caused by crop planting types, it is difficult for traditional distribution network to meet the electricity demand in irrigation period. In order to improve the reliability of agricultural irrigation, this paper adopts L-ISODATA algorithm to analyze the agricultural irrigation load data, aiming at identifying different agricultural irrigation load modes. It provides a basis for the optimal allocation and management of distribution network resources. In this paper, L-ISODATA clustering algorithm is used to cluster the processed agricultural irrigation load data set. Then, by comparing different clustering effects, the characteristics of agricultural irrigation load in different regions and different periods were analyzed. Finally, through cluster analysis, the agricultural irrigation load curve is divided into several representative categories, each category reflects a specific irrigation demand model. The findings of this paper have important implications for understanding the time distribution characteristics of agricultural irrigation, guiding the reasonable allocation of distribution resources and improving the utilization efficiency of distribution resources.