Computational molecular design (CMD) is a methodology which applies optimization techniques to develop novel lead compounds for a variety of applications. In this work, a CMD method is applied to the design of ionic liquids (ILs), which are being considered for use as environmentally benign solvents. The molecularly tunable nature of ILs yields an extraordinary number of possible cation and anion combinations, the majority of which have never been synthesized. The product design framework developed in this work seeks to accelerate the commonly used experimental trial-and-error approach by searching through this large molecular space and providing a set of chemical structures likely to match a set of desired property targets. To predict the physical and chemical properties of an ionic liquid in a specific system, quantitative structure-property relations (QSPRs) have been developed. In this work, correlations were created for solubility, diffusivity, and melting temperature. The electronic structure of ionic liquids is quantified using molecular connectivity indices, which describe bonding environments, charge distribution, orbital hybridization and other interactions within and between ions. The resulting property prediction model is then integrated within a computational molecular design framework, which combines the QSPRs with structural feasibility constraints in a combinatorial optimization problem. The problem is reformulated as an MILP after exact linearization of structural constraints. An example is provided to test the formulation for the design of ionic liquids for use within a hydrofluorocarbon (refrigerant) gas separation system. A second example compares a stochastic optimization algorithm, Tabu Search, to a standard deterministic solver for the solution of a larger-scale refrigeration design problem. The computational efficiency and practical implementation of this product design methodology is also discussed. (C) 2010 Elsevier Ltd. All rights reserved.