AbstractHow to assess the temperature sensitivity (Q10) of soil organic matter (SOM) decomposition and its regional variation with high accuracy is one of the largest uncertainties in determining the intensity and direction of the global carbon (C) cycle in response to climate change. In this study, we collected a series of soils from 22 forest sites and 30 grassland sites across China to explore regional variation inQ10and its underlying mechanisms. We conducted a novel incubation experiment with periodically changing temperature (5–30 °C), while continuously measuring soil microbial respiration rates. The results showed thatQ10varied significantly across different ecosystems, ranging from 1.16 to 3.19 (mean 1.63).Q10was ordered as follows: alpine grasslands (2.01) > temperate grasslands (1.81) > tropical forests (1.59) > temperate forests (1.55) > subtropical forests (1.52). TheQ10of grasslands (1.90) was significantly higher than that of forests (1.54). Furthermore,Q10significantly increased with increasing altitude and decreased with increasing longitude. Environmental variables and substrate properties together explained 52% of total variation inQ10across all sites. Overall,pHand soil electrical conductivity primarily explained spatial variation inQ10. The general negative relationships betweenQ10and substrate quality among all ecosystem types supported the C quality temperature (CQT) hypothesis at a large scale, which indicated that soils with low quality should have higher temperature sensitivity. Furthermore, alpine grasslands, which had the highestQ10, were predicted to be more sensitive to climate change under the scenario of global warming.