Revealing the relationships between the temperature sensitivity of soil organic matter decomposition (Q10) and microbial communities is critical to predict ecosystem feedbacks to global warming. However, the relationships are still far from well understood, especially within the low latitude regions. To address this knowledge gap, soil samples were collected from >100 cropland sites in Southwest China, a region with highly diverse climatic and environmental conditions. The results showed that Q10 values substantially varied across the region, ranging from 1.85 to 6.81, with an average value of 3.27. They were significantly positively correlated with the contents of soil organic carbon, while negatively correlated with the ratios of soil dissolved organic carbon to total organic carbon content. This indicates that soils with high organic carbon contents are more vulnerable to global warming. Further analysis suggested that Q10 values were positively correlated with soil microbial respiration rates, fungal abundance, prokaryotic diversity, and the relative abundance of copiotrophic microbial lineages, but negatively correlated with the proportions of oligotrophic microbes and microbial co-occurrence network degree. The structural equation modeling analysis suggested that soil organic carbon and its quality, as well as microbial attributes were the main factors explaining the variation in Q10 values. The findings in this study highlight the crucial and complex roles of soil microbiome in determining ecosystem feedbacks to global warming.