Light use efficiency (LUE) is a key parameter in terrestrial carbon cycling, yet its spatial variability cannot be fully explained by climatic factors alone. Plant functional traits are expected to provide important physiological constraints on vegetation photosynthesis, but their roles and interactions with climatic drivers remain poorly understood, limiting the accuracy and improvement of carbon cycle modelling and projections. Here, we use eddy covariance observations from 29 flux sites, together with climate and trait data, to identify the drivers of LUE and map its spatiotemporal variations across China. Incorporating plant functional traits, including rooting depth, isohydricity, leaf nitrogen (N) content and specific leaf area (SLA), substantially improves the explanation of LUE spatial variability from 45 % to 58 %. The attribution analysis reveals that this improvement is mainly contributed by the nonlinear interactions between climatic factors and traits, accounting for 69.1 % of the total importance and considerable exceeding the direct contributions from individual factors. Typically, the SLA-related interactions, particularly those involving rooting depth, temperature and vapor pressure deficit, largely explain tower-based LUE variations in China, suggesting that LUE is regulated by the coordination between vegetation carbon acquisition, soil water access and atmospheric constraints. Based on these findings, we generate a 1-km resolution annual LUE dataset for China spanning from 2003 to 2024. The regional LUE exhibits pronounced spatial heterogeneity but shows non-significant long-term trends during the study period. Our results emphasize that vegetation LUE distributions are mainly governed by the coupled climate–trait interactions, highlighting the need to represent trait-mediated environmental responses in terrestrial carbon cycle modelling.