AbstractA conceptual model based on the surface energy budget is developed to compute the sensitivity of the climatological mean diurnal amplitude and mean daily surface temperature to the soil thermal inertia. It uses the diurnal amplitude of the net surface radiation, the sensitivity of the turbulent fluxes to the surface temperature and the soil thermal inertia. Its performance is evaluated globally with numerical simulations using the atmospheric and the land surface modules of a state‐of‐the‐art climate model. The only regions where the thermal inertia has a limited impact are the moist regions. In dry areas, together with the stability of the boundary layer it plays a major role. It also has a significant impact at high and midlatitudes especially in winter when the turbulent fluxes are weak. In semiarid regions, the soil moisture exhibits a high day‐to‐day variability strongly correlated to the evaporation and the day‐to‐day variability of the surface temperature is generally explained by the soil moisture via its control on the evaporation. However, the soil moisture via its control on the thermal inertia reduces the impact of the day‐time evaporative cooling by reducing the nocturnal cooling. This newly identified moisture‐related negative feedback can reduce the variability of the surface temperature in semiarid regions by up to a factor of 2. The model also provides a simple framework to understand the role of the thermal properties in the frequent cold bias identified in stratified stable atmospheric situations in the northern midlatitudes.