Improving irrigation efficiency is essential for sustaining winter wheat production in water-limited agricultural systems. This study developed a high-resolution remote sensing framework to quantify actual crop evapotranspiration (ETa), irrigation status, yield, and water use efficiency (WUE) across three winter wheat growing seasons (2018-2019, 2019-2020, and 2020-2021) over the Punjab, Pakistan region. Sentinel-2 imagery was integrated with the Simple Algorithm for Evapotranspiration Retrieving (SAFER) to derive land surface temperature (LST), albedo, normalized difference vegetation index (NDVI), crop coefficient (Kc), and ETa at 10 m spatial resolution, while ERA5-Land data provided meteorological forcing. A radiation-based biomass model was applied to estimate yield, and WUE was calculated as the ratio of yield to ETa. Validation was conducted using survey-based observations collected from 50 farmers and field boundaries derived from cadastral records. The results revealed clear seasonal and spatial dynamics in crop condition, water consumption, and productivity. All variables followed a coherent phenological pattern, with low values during crop establishment, peak canopy development and water use during January-March, and rapid decline during senescence. Marked field-scale heterogeneity was observed in ETa, irrigation status, yield, and WUE, indicating substantial variation in irrigation behaviour and crop performance within the study area. Notably, irrigation status analysis indicated peak over-irrigation in January-March, highlighting the need for optimized mid-season irrigation scheduling. Among the three seasons, 2018-2019 exhibited comparatively stronger ETa and WUE, whereas 2020-2021 showed reduced peak performance. The validation results demonstrated a strong agreement between observed and simulated outputs. For yield, R2 values were 0.74, 0.63, and 0.60, with corresponding RMSE values of 126.75, 131.92, and 155 kg/acre for the years 2018-2019, 2019-2020, and 2020-2021, respectively. For Water Use Efficiency (WUE), R2 values were 0.78, 0.69, and 0.63, with RMSE values of approximately 0.10, 0.09, and 0.09 kg/m3, respectively, indicating a moderate to strong correlation between observed and simulated values across the three years. Overall, the proposed framework provides a robust and scalable basis for evaluating irrigation performance and improving field-scale water productivity in semi-arid agricultural regions.