The ongoing increase of atmospheric temperature may induce soil organic carbon (SOC) loss and exacerbate the greenhouse effect. As a result, there is a great effort to understand the relationship between temperature and the heterotrophic soil respiration rate (RSOIL) as it has significant implications for anticipated change of the Earth system. Soil respiration depends on the size of respiring microbial biomass (MBC) and when RSOIL is measured without concurrent measurement of MBC, the apparent temperature sensitivity of RSOIL could be misinterpreted since MBC can change with temperature within days or weeks of warming. The effect of temperature driven changes in MBC on the apparent temperature sensitivity of RSOIL was evaluated using a meta-analysis of 27 laboratory and field experiments conducted at different temporal scales (1–730 d) and under a wide range of temperatures (2–50 °C) and soil conditions. Across all studies, the apparent temperature sensitivity decreased when MBC decreased with increasing temperature and vice versa. We observed a steep decrease of MBC above optimal temperature (27.1 ± 1.0 °C), which attenuated the apparent temperature sensitivity of RSOIL, an aspect previously explained by the existence of reaction rate temperature optima. The temperature response of the MBC specific respiration rate was, however, highly non-linear and soil specific. Including MBC in soil biogeochemical models requires careful consideration of the variability of temperature-associated physiological changes of soil microorganisms. Without it, microbially explicit models cannot predict temperature induced SOC loss better than older, empirical models based on first order reaction kinetics.