The expanding applications of ionic liquids (ILs) give rise to a need to combine their efficiency with their environmental and economic sustainability. Experimental investigations can explore only a tiny portion of the enormous chemical space covered by IL cation and anion combinations. Hence the utility of the recently derived in silico VolSurf+ descriptors to develop quantitative structure–property relationships (QSPRs). In this context, multivariate data-driven approaches taking into account the simultaneous variation of both anionic and cationic counterparts of ILs turn out to be suitable to predict ILs toxicities as well as physico-chemical properties by means of partial least squares (PLS) modelling. Examples of applications for the design of sustainable and efficient ILs are reported and the potentialities of such multivariate approaches for smart ILs and materials design are pointed out.