Ionic liquids (IL) are considered viable alternative solvents for CO2 capture. Because of this and the unlimited number of possible combinations between cations and anions, it is important to develop a systematic approach based on a reliable model for defining the best combinations to promote natural gas sweetening by enabling the most efficient CO2 capture accompanied by minimal hydrocarbons losses. Hence, in this study, the conductor-like screening model for real solvents (COSMO-RS) model was employed to predict the solubility of CO2, CH4, C2H6, and C3H8 in eight selected ILs with varying combinations of cations and anions. All predictions have been compared against experimental data from the literature which was obtained at the same conditions. The model and experimental data show similar trends and moderate agreement for each of the selected ILs. The observed deviations were 8–62%, 8–40%, 1–37%, and 29–35%, for CO2, CH4, C2H6, and C3H8, respectively. Both the experimental data and predictions confirm that [hmim][TCB] IL is more suitable for natural gas sweetening because of its higher CO2 solubility and selectivity than that of other ILs. Finally, the experimentally determined solubilities of these gases in ILs and conventional amine solvents are compared with highlight the greater efficiency of ILs.