Due to the intrinsic nature of a heterogeneous and anisotropic microstructure of cement based materials and the small indentation size, the measured microhardness values are subject to considerable variability. This work presents an in-depth assessment of microhardness of hardened cement pastes (HCPs) from a statistical perspective. Hundreds of microhardness measurements were conducted on an HCP sample using a Vickers Microhardness Tester. The results showed that the microhardness measured from the HCP sample significantly scattered with a large standard deviation, varying from tens to hundreds. The data-set of microhardness values was not normally distributed but fit best with a three-parameter lognormal model. By using a statistical software, the probability density function of the microhardness distribution can be readily obtained. The arithmetic mean and its 95％ confidence intervals of the measured microhardness values can be used to best represent the microhardness characteristics of HCPs.