Nuclear Magnetic Resonance (NMR) is an important tool for the evaluation of pore structure and petrophysical properties of reservoir rocks. Pore Surface relaxivity (ρ2), which controls the relaxation decay due to the interaction between the spins and the pore surface, is an important parameter that impacts the interpretation of NMR for petrophysical properties. Sandstone rocks have complex minerology and consist of different clay minerals. Earlier efforts have reported the surface relaxivity of different rock types such as sandstone or carbonate, but less attention was given to the estimation of individual clay minerals surface relaxivity. The surface complexity and variable mineralogy (especially paramagnetic and ferromagnetic minerals) of clays can both impact the surface relaxivity and relaxation times in clays and porous media containing clays. This study, using for the first time both simulation and experimental approaches, aims to systematically investigate the surface relaxivity of clays characterized by variable mineralogy and iron content. A glass beads sample (iron free) and 5 different pure clay minerals samples (ranked by increasing iron content: Kaolinite, Smectite, Illite, Chlorite, and Nontronite) were prepared with similar average grain sizes (500 μm). The determination of surface relaxivity requires prior measurement of pore surface-to-volume ratio (S/V) using an independent method in addition to obtaining T2 relaxation time. Two methods, among the most widely used, were used here for S/V determination including BET N2 gas adsorption and 3D micro CT imaging. The 3D images were also used to conduct NMR simulations to reproduce the experimental data. The estimated surface relaxivity values from both methods were compared and correlated to the iron content of the different clays. The surface relaxivity values obtained based on BET-derived S/V showed reasonable values (29–343 μm/s) that have a strong positive relationship with iron content of the clays. The documented equation relating surface relaxivity and iron content can be potentially utilized to predict surface relaxivity based on the compositional analysis of clays. On the other hand, digital images did not capture the surface complexity of clays thus significantly underestimated S/V which consequently resulted in overestimated surface relaxivity (898–11000 μm/s) that were physically non-feasible. Supporting the same argument, the simulated T2 relaxation distribution showed significant mismatch with the experimental data for the clay samples while an excellent match was observed for the glass beads sample. This can be attributed to the smooth surface of the glass beads, unlike the complex clays. The reported surface relaxivity values (based on BET-derived S/V) for the individual clay minerals can be used to perform NMR simulation and interpretation studies conducted on sandstone or shale formations. Careful attention should be given when performing NMR simulation on micro-CT digital images of samples containing clay due to the inadequacy of such images in capturing the complex surface and microporous nature of clays. The outcomes from this study can improve the interpretation of NMR T2 data and the determination of NMR-derived pore size distribution and petrophysical properties. The reported surface relaxivity values for clay can also inform the NMR simulation.