Temperature is a major environmental variable influencing microbial respiration in soils. Thus, understanding how heterotrophic processes in soils may respond to potential increases in temperature is crucial for the prediction of the response of forest carbon budgets to climate change. We investigated carbon mineralization rates from eight European forest soils in relation to soil temperature. Mineral soil samples were collected from eight mature forest sites in the European network CARBOEUROFLUX and were incubated in the laboratory for ca. 270 days at four temperatures: 4, 10, 20 and 30°C. In all soils, carbon mineralization rates decreased over time when incubated at high temperatures of 20 and 30°C. In this study, we explore the different models available to analyse long-term incubation data. Carbon mineralization rates were best predicted by a first-order, two-compartment model that predicted carbon mineralization as a function of time and temperature using all of the incubation data. We found very small fractions (1–9%) of labile carbon in the upper mineral soils. Despite large differences among sites, we found higher carbon mineralization rates and larger amounts of labile carbon in the broadleaf than in the conifer forest soils. No significant differences in temperature sensitivity among the sites (average Q10 of 2.88 using the two-compartment model) were observed, as estimated with all methods used. Although not statistically significant, the sensitivities of the rate constant of the labile fractions tended to be higher than those for the rate constant of the recalcitrant fractions. Thus, the results of this modelling exercise suggest that despite large variation among sites, a single temperature sensitivity parameter can be used for a range of soils over the range of temperatures we used (4–30°C).