Understanding how biodiversity sustains ecosystem functioning is fundamental both in theory and practice, particularly in the context of primary productivity, which underpins many forms of life and ecosystem processes such as carbon sequestration. Using an extensive dataset from Japan’s National Forest Inventory—spanning over two decades and 6.4 million tree measurements across subtropical to hemiboreal forests—we examined the relationship between tree diversity and aboveground biomass productivity (AGBP) in natural forests. We found that species richness positively influences AGBP, primarily through the positive complementarity effect, with marginal contributions from the negative selection effect. By integrating broad-scale environmental niche dissimilarity and local-scale functional differentiation in light of the metacommunity theory, we quantified diversity effects and predicted productivity outcomes under forest restoration scenarios. Our models suggest that enhancing tree diversity in unmanaged, abandoned plantations could enhance aboveground biomass productivity up to 7.6 times compared to the current baseline. This evidence underscores the irreplaceable value of naturally-assembled communities and highlights the necessity of species-explicit practices, as failing to consider species composition may limit ecosystem function recovery. Our study provides critical insights for biodiversity-based climate solutions and sustainable forest management policies.