<p>Accurate estimation of evapotranspiration (ET) via remote sensing is crucial for understanding the hydrological cycle, enhancing water use efficiency, and effectively managing irrigation under water scarcity. The thermal-optical trapezoidal model is widely used for estimating regional ET using remotely sensed data. However, this method limits ET estimation to thermal sensors with longer revisiting times and lower spatial resolution. Furthermore, it creates significant uncertainty due to the necessity for extensive spatial and temporal calibration. To address these limitations, this study proposes an automated approach using the novel Optical Trapezoid Model for crop evapotranspiration (OPTRAM-ETc), which leverages high-resolution optical data from Sentinel-2 to generate field-scale ETc. The method parameterizes the shortwave-infrared transformed reflectance (STR)–vegetation index (VI) trapezoid and automatically derives wet and dry edges, thereby minimizing user intervention and avoiding site-specific calibration. Applied to sugarcane cultivation systems in Khuzestan, Iran, over three consecutive growing seasons (2018–2021), the proposed approach demonstrated robust performance in estimating ETc, achieving R² values between 0.87 and 0.89 and RMSE values of 1.64–1.99 (mm d^-1) across the three crop years. Incorporating multi-year temporal fusion further enhanced model performance (R² = 0.91), demonstrating temporal consistency and robustness. In addition, various VIs (Normalized Difference Vegetation Index (NDVI), Fractional Vegetation Cover (FVC), Soil-Adjusted Vegetation Index (SAVI), Enhanced Vegetation Index (EVI), Red Edge Normalized Difference Vegetation Index (RENDVI), and Modified Beer-Lambert Law (MBLL)) were employed in the OPTRAM-ETc model, and the results indicated that the choice of vegetation index did not significantly impact the accuracy of ETc estimation, confirming the model’s generalizability. The high-resolution ETc outputs enable spatially explicit irrigation scheduling, improve early-season water stress detection, and support long-term monitoring. These advances provide actionable insights for water managers and policymakers in water-scarce regions, directly addressing complex environmental challenges in agricultural and hydrological systems.</p> Graphical Abstract <p>This graphical abstract illustrates the workflow and main findings of the proposed OPTRAM-ETc framework for estimating crop evapotranspiration (ETc) using high-resolution Sentinel-2 observations. The study focuses on irrigated sugarcane fields in Khuzestan, Iran, over three consecutive growing seasons (2018–2021). First, the locally calibrated FAO-56 Penman–Monteith model (LCETc) was used, incorporating meteorological variables together with field-measured sugarcane height at each point to establish a ground-based reference for ETc. Subsequently, Sentinel-2 imagery was processed on the Google Earth Engine platform to extract various vegetation indices (VIs; NDVI, NDWI, EVI, MBLL, SAVI, RENDVI) along with shortwave infrared transformed reflectance (STR). A two-step smoothing procedure was applied to minimize temporal noise. The core innovation of the OPTRAM-ETc approach lies in parameterizing the STR–VI trapezoidal space. Automatic detection of wet and dry edges was carried out using two parameterization techniques, the Interval-Based Regression and Filtering (IRF) and the Density-Enhanced Interval-Based Regression (D-IRF), within the Python environment, thereby reducing the need for site-specific calibration and minimizing user intervention. Model evaluation across three growing seasons (2018–2021) confirmed robust accuracy, which was further enhanced through multi-year temporal fusion (R² = 0.91). Overall, the framework produces field-scale, high-resolution ETc maps, offering practical tools for irrigation scheduling, early water stress detection, and long-term agricultural water management.</p>