Soils harbor more than three times as much carbon (C) as the atmosphere, a large fraction of which (stable organic matter) serves as the most important global C reservoir due to its long residence time. Litter and root inputs bring fresh organic matter (FOM) into the soil and accelerate the turnover of stable C pools, and this phenomenon is termed the “priming effect” (PE). Compared with knowledge about labile soil C pools, very little is known about the vulnerability of stable C to priming. Using two soils that substantially differed in age (500 and 5300？years before present) and in the degree of chemical recalcitrance and physical protection of soil organic matter (SOM), we showed that leaf litter amendment primed 264% more organic C from the young SOM than from the old soil with very stable C. Hierarchical partitioning analysis confirmed that SOM stability, reflected mainly by available C and aggregate protection of SOM, is the most important predictor of leaf litter-induced PE. The addition of complex FOM (i.e., leaf litter) caused a higher bacterial oligotroph/copiotroph (K-/r-strategists) ratio, leading to a PE that was 583% and 126% greater than when simple FOM (i.e., glucose) was added to the young and old soils, respectively. This implies that the PE intensity depends on the chemical similarity between the primer (here FOM) and SOM. Nitrogen (N) mining existed when N and simple FOM were added (i.e., Glucose+N), and N addition raised the leaf litter-induced PE in the old soil that had low N availability, which was well explained by the microbial stoichiometry. In conclusion, the PE induced by FOM inputs strongly decreases with increasing SOM stability. However, the contribution of stable SOM to CO2 efflux cannot be disregarded due to its huge pool size.