Emulsion liquid membrane (ELM) is an effective way to recover and remove biologically active compounds, particularly drug molecules from hospital wastewater. Ionic liquid (IL)-assisted ELMs have recently been found to be efficient for recovery and removal of biologically active compounds owing to their excellent characteristics. However, the selection of potential ILs from the wide range of anion-cation combinations is quite challenging. To address this issue, the conductor-like screening model for real solvents (COSMO-RS) has emerged as a reliable tool to screen solvents based on a variety of fluid mixture properties. In this study, we screened ILs for the extraction of lactic acid as a model drug using COSMO-RS. Six cations and 20 anions were selected and combined to form 120 ILs, which were then assessed for recovery of lactic acid. The activity coefficient at infinite dilution, capacity, selectivity, performance index, and solvation energy values for all of the combinations were predicted, by which the effectiveness of each IL was evaluated. ILs without aromatic rings showed greater efficiencies than ILs containing aromatic rings. Moreover, the sulphate anion-cation combinations showed the highest capacities and selectivities. Overall, tetramethylammonium sulfate was found to be the best extractant for lactic acid. The COSMO-RS predictions were validated with experimental data, and the results showed a maximum difference of ~8.0 between the predicted and experimental values. © 2020 Elsevier B.V.