<p>Understanding how human-altered landscapes influence evapotranspiration (ET) remains a major challenge in hydrological modelling. This study develops an integrated framework combining Airborne Laser Scanning (ALS) and the Global Land Data Assimilation System (GLDAS) to quantify the effects of vegetation structure and topographic variability on ET in contrasting anthropogenic environments of northern and central Poland. ALS point clouds were processed to derive vegetation fraction and slope coefficient, which were then incorporated into the GLDAS-based ET estimates through a physically informed correction scheme. The modified approach enables differentiation between industrial, urban, and agricultural areas, linking their morphological characteristics with water and energy exchange processes. Results demonstrate that dense vegetation significantly suppresses ET amplitude by as much as eightfold compared to sparsely vegetated agricultural landscapes, while terrain slope exerts a secondary but measurable damping effect. Despite these spatial contrasts, the temporal synchrony of ET among all sites reveals strong regional climatic control. The method thus provides a scalable means to improve ET representation in hydrological models by explicitly accounting for canopy and terrain influences derived from high-resolution ALS data. This approach strengthens process understanding of water-energy feedback in human-modified environments and contributes to improved coupling between remote-sensing observations and physically based hydrological modelling frameworks. Quantitatively, ET amplitude decreased by approximately fourfold in the industrial, eightfold in the agricultural area, and by about 50% in the urban landscape after ALS-based correction. These reductions, closely linked to vegetation fraction (ranging from 11% to 53%) and terrain slope (4.66–29.29%), highlight the dominant role of canopy structure in regulating ET dynamics across diverse human-modified terrains.</p>