In the context of water scarcity, understanding the mechanisms influencing and altering agricultural water consumption can offer valuable insights into the scientific management of limited water resources. Using Henan Province as a case study, this research applies the Mann–Kendall test method, the spatial Markov transfer chain model, the optimal parameter geo-detector model, and the Logarithmic Mean Divisia Index (LMDI) decomposition method to investigate the evolution characteristics of agricultural water consumption in Henan Province and its key influencing factors. The findings revealed the following: (1) Agricultural water consumption has shown a significant decline from 1999 to 2022. (2) According to observations, the stability of agricultural water consumption exceeds the spillover effect, and cross-border grade transfer is challenging. Moreover, this phenomenon is influenced by the neighboring regions. (3) The key influencing factors of added agricultural value are the sown area of food crops, total sown area, irrigated area, and average annual air temperature. (4) Among the decomposition effects on agricultural water consumption, the contribution of each decomposition effect to changes in agricultural water consumption and the role of spatial distribution exhibit notable differences. Overall, these findings provide theoretical references for the efficient use of agricultural water resources and sustainable development in the region.