The concept of short time scale mean-square displacement (MSD) was used to get a picture of the distribution of atomic mobilities in different layers and regions of Au and Pt nanoparticles (NPs). The NPs were simulated in vacuum at different temperatures using molecular dynamics and the embedded-atom model. The calculated atomic mobilities were greater for atoms located at corner positions, followed by atoms on edges and planes, independently of the layer analyzed. The short time scale MSD was revealed to be an excellent alternative to predict melting temperatures of small (<100 atoms) transition-metal NPs. Finally, the combination of classical (MSD) and quantum (density of states) properties brought some insight into how the catalytic activity may locally change over the NP surface. The trends found for subnanometer-sized NPs indicate that corner, kinks, and defect regions play a major role for the catalytic activity of these particles.