Global change impacts on biodiversity are highly context-dependent, complicating predictions of ecosystem functioning. We synthesized 18,739 control-treatment paired observations from 680 studies to assess plant diversity responses to simulated global change, identify key mechanisms, and evaluate associated ecological consequences. Our results show that soil texture (sand-to-silt ratio) is the dominant spatial predictor of plant diversity. Elevated CO2 had a neutral effect on species richness, whereas increased precipitation enhanced it. In contrast, warming, drought, and nitrogen addition reduced species richness, with nitrogen addition consistently suppressing Shannon diversity and evenness. Nearly half of all two-driver interactions were additive, although non-additive effects were also common. Contextual factors, including initial diversity, sampling size, experimental duration, and soil properties, mediated plant diversity dynamics under global change. Importantly, shifts in plant diversity altered ecosystem productivity, carbon and water fluxes, and soil biota, underscoring the need to incorporate biodiversity and site-specific context into global change-ecosystem feedback models.