Abstract Temperature sensitivity (Q10) of soil organic matter (SOM) releases is a crucial parameter in land surface models for predicting the fate of permafrost carbon (C) under climate warming. Despite its importance, it remains largely unknown whether warming scenarios may cause heterogeneous responses of SOM decomposition in high-latitude and high-altitude permafrost regions. Here, we conduct a 368-day laboratory incubation experiment using topsoil samples collected from 30 sites across high-latitude (Northeast China) and high-altitude (Tibetan Plateau) permafrost regions to elucidate Q10 patterns and their drivers. We find that the Q10 value in the high-latitude permafrost regions is ∼1.8 times higher than that in the high-altitude permafrost regions. This finding is further supported by a literature synthesis of Q10 data from the Arctic permafrost regions. Soil microbial properties (that is, microbial composition, α diversity, life-history strategies, and C-degrading genes) and mineral protection exert the strongest direct effects on Q10 variation, whereas climate (aridity), soil characteristics (pH and clay + silt), and substrate quantity influence Q10 indirectly across the two permafrost regions. These findings highlight that permafrost regions at high latitudes, characterized by higher Q10 values and substrate quantity, may be more susceptible to C losses under warming scenarios compared with high-altitude permafrost regions.