Understanding how soil microorganisms influence the direction and magnitude of soil carbon feedback to global warming is vital to predict future climate change. Although microbial activities are major contributors to soil respiration (RS) and its temperature sensitivity (Q10), the mechanisms underpinning microbial influence on RS and Q10 remain unclear. In this study, structural equation modeling (SEM) was conducted to illustrate that bacteria mainly affect RS by shifting beta diversity (denoted NMDS ordinations) instead of richness. In contrast, Q10 values are governed by the richness and NMDS ordinations of bacteria. We also found that soil water content (SWC) was the factor key to changing bacterial properties (P < 0.05, R2 ≥ 0.33). Network analysis demonstrated that only Proteobacteria were positively associated with RS (P < 0.05, R > 0.5). Illuminating the mechanisms underpinning the influence of soil microbes on RS and Q10 values is fundamental to understanding mechanistic soil-climate carbon cycles.