A major challenge of quantifying feedback between microbial communities and climate is the vast diversity of microbial communities and the intricacy of soil biogeochemical processes they mediate. We overcome this challenge by simplifying the representation of diverse enzyme functions from metagenomics data. We developed a dynamic allocation scheme for enzyme functional classes (EFCs) based on the premise that microbial communities act to maximize acquisition of limiting resources while minimizing energy expenditure for acquiring unlimited resources. We incorporated this scheme into a biogeochemical model to explicitly represent microbial functional diversity and simulate responses of microbially-mediated soil biogeochemical processes to varying environmental and nutrient conditions. Representing microbial functional diversity and environmental acclimation improved predictions of the stoichiometry of microbial biomass and mitigated the sensitivity of soil organic carbon to warming in nutrient-deficient regions. Our results indicate the importance of microbial functional diversity and environmental acclimation for projecting climate feedbacks of nutrient-limited soils.