Although ionic liquids (ILs) have been widely explored as solvents for extractive desulfurization (EDS) of fuel oils, systematic studying of the optimal design of ILs for this process is still scarce. The UNIFAC-IL model is extended first to describe the EDS system based on exhaustive experimental data. Then, based on the obtained UNIFAC-IL model and group contribution models for predicting the melting point and viscosity of ILs, a mixed-integer nonlinear programming (MINLP) problem is formulated for the purpose of computer-aided ionic liquid design (CAILD). The MINLP problem is solved to optimize the liquid-liquid extraction performance of ILs in a given multicomponent model EDS system, under consideration of constraints regarding the IL structure, thermodynamic and physical properties. The top five IL candidates preidentified from CAILD are further evaluated by means of process simulation using ASPEN Plus. Thereby, [C5MPy][C(CN)(3)] is identified as the most suitable solvent for EDS. (c) 2017 American Institute of Chemical Engineers