Abstract: Accurate identification of irrigated farmland is crucial for optimizing water resource allocation, improving irrigation efficiency, monitoring agricultural drought, guiding hydraulic engineering operations, and promoting food security and sustainable water use. This study focuses on Uxin Banner, Inner Mongolia, located in an arid–semiarid agro-pastoral ecotone. Differences between irrigated and non-irrigated farmland were comparatively analyzed from four dimensions—moisture dynamics, energy dynamics, vegetation physiology, and vegetation phenology—leading to the selection of 10 key factors across four categories. These factors were retrieved using multi-source remote sensing data to construct a comprehensive feature index library containing 43 indicators. The optimal feature subset was selected through feature-importance evaluation, and a Random Forest (RF) model for irrigated farmland identification was trained based on this subset. The model was applied to identify irrigated farmland in Uxin Banner from 2019 to 2024 and to analyze its spatiotemporal variations. Results show that the model achieved an overall accuracy of 95.38% with a Kappa coefficient of 0.9128, demonstrating strong capability in irrigated farmland extraction. In 2024, the irrigated farmland area in Uxin Banner was 812.2733 km2, mainly distributed in the southern and southwestern regions. Over the past five years, irrigated farmland area exhibited a turning point in 2022, increasing annually before 2022 and declining thereafter. The proposed framework—mechanism analysis, factor identification, remote sensing inversion, and model identification—provides a reference for irrigated farmland mapping and interpretable machine learning applications.