Plant diversity enhances many ecosystem functions, including root biomass production, which drives soil carbon input. Although root decomposition accounts for a large proportion of carbon input for soil, little is known about plant diversity effect on this process. Plant diversity may affect root decomposition in two non-exclusive ways: by providing roots of different substrate quality (e.g. root chemistry) and/or by altering the soil environment (e.g. microclimate). To disentangle these two pathways, we conducted three decomposition experiments using a litter-bag approach in a grassland biodiversity experiment. We hypothesized that: (i) plant species richness negatively affects substrate quality (indicated by increased C:N ratios), which we tested by decomposing roots collected from each experimental plot in one common plot; (ii) plant species richness positively affects soil environment (indicated by increased soil water content), which we tested by decomposing standardized roots in all experimental plots; (iii) the overall effect of plant species richness on root decomposition, due to the contrast between quality and environmental effects, is neutral, which we tested by decomposing community roots in their ‘home’ plots. Plant species richness negatively affected root decomposition in all three experiments. The negative effect of plant species richness on substrate quality was largely explained by increased root C:N ratios along the diversity gradient. Functional group presence explained more variance in substrate quality than species richness. Here, the presence of grasses negatively affected substrate quality and root C:N ratios, while the presence of legumes and small herbs had positive effects. Plant species richness had a negative effect on soil environment despite its positive effect on soil water content which is known to stimulate decomposition. We argue that – instead of soil water content – a combined effect of soil temperature and seasonality might drive environmental effect of plant diversity on decomposition in our plant communities, but this remains to be tested. Synthesis. Our results demonstrate that both substrate quality and soil environment contribute to the net negative effect of plant diversity on root decomposition. This study promotes our mechanistic understanding of increased soil carbon accumulation in more diverse grassland plant communities.