Recently, water shortage has increasingly threatened agricultural production and ecological health in many parts of the world. Water is scarce in regions like Debrecen, Hungary, where climate change exacerbates the situation. This study explores the use of satellite data in crop classification for water management, analyzing how crop variety influences water consumption and evapotranspiration (ET), with a focus on crop selection to achieve sustainable water use. The research examines the 2022 drought and how crop classification, combined with a machine learning technique can identify water-intensive crops and encourage efficient water use. A Random Forest, Classification And Regression Trees, and Support vector machine model were trained, and its accuracy was evaluated using a confusion matrix and other metrics. The land cover map and 2021 ground truth data showed a higher overall accuracy of the Random Forest model 95%, with a Kappa coefficient of 92%, indicating the model performed exceptionally well. Based on these results, the model was applied to the years 2019, 2020, 2021, 2022, 2023, and 2024. Area coverage for each plant species was calculated, enabling us to identify dominant crops and estimate the water balance requirements of each crop in the region.