The abundance of aqueous ions plays a significant role in various processes, including environmental ones such as cloud condensation nucleation. The details of ion-water and water-water interactions are critically important in these processes and require further insights from molecular simulations. Towards elucidating the hydration of ions in clusters, molecular dynamics (MD) simulations based on first-principles, implementing the density-functional tight-binding model with the third-order expansion (DFTB3), were employed to predict the properties of water clusters of varying sizes seeded with polyatomic ions (ClOx− and NOx−). Validating the predictions of the DFTB3 model against high-level quantum chemistry data established the robustness and accuracy of DFTB3 in evaluating small cluster properties of ClOx− (H2O)n and NOx− (H2O)n. In order to enhance conformational sampling of the clusters, a parallel tempering approach of MD simulations was implemented, and the potential of mean force (PMF) along the ion to water cluster center-of-mass distance was calculated by implementing umbrella sampling in the DFTB3-MD simulations. On the basis of the probability distributions of the ions with respect to the water cluster moiety center of mass obtained from the PMF, it appears that ClOx− and NOx− ions exhibit surface behaviour in water clusters, with the larger oxides showing more preference for surface solvation. This feature may influence the chemistry of the ions as they will be readily available for reactions with other atmospheric particles.