The formation of thin oxide films on metal supports is an important phenomenon, especially in the context of strong metal support interaction (SMSI). Computational predictions of the stability of these films are hampered by their structural complexity and a varying lattice mismatch with different supports. In this study, we report a large combination of supports and ultrathin oxide films studied with density functional theory (DFT). Trends in stability are investigated through a descriptor-based analysis. Since the studied films are bound to the support exclusively through metal–metal interaction, the adsorption energy of the oxide-constituting metal atom can be expected to be a reasonable descriptor for the stability of the overlayers. If the same supercell is used for all supports, the overlayers experience different amounts of stress. Using supercells with small lattice mismatch for each system leads to significantly improved scaling relations for the stability of the overlayers. This approach works well for the studied systems and therefore allows the descriptor-based exploration of the thermodynamic stability of supported thin oxide layers.