Maximum light use efficiency (εmax) represents a plant's capacity to convert light into carbon during photosynthesis. Although prior studies have explored εmax variations between sunlit and shaded leaves or its temporal ties to canopy structure, the spatial relationship between biome-level εmax (εbiome) and biome structure remains poorly understood. We analysed data from 320 eddy covariance sites ( 855 site-years) with satellite-derived near-infrared reflectance of vegetation (NIRv) and leaf area index (LAI). We introduced NIRvN (NIRv/LAI) to isolate architectural effects from leaf quantity. Site-level εmax was calculated and aggregated by biome to derive εbiome. Results show εbiome rises nonlinearly with NIRv and LAI, saturating at high LAI, with crops and tropical evergreen forests deviating from this trend. Conversely, εbiome decreases linearly with increasing NIRvN, indicating that biomes with greater NIR scattering efficiency exhibit lower εbiome. These results enhance understanding of structural influences on carbon uptake across global biomes.