In this study, a new accurate model is presented for estimation of the glass transition temperature of ionic liquids. To develop this simple model, the contribution of ILs' anions and cations are separately considered through a number of occurrences of 50 substructures (17 anion-based and 33 cation-based). This simple model shows a low average relative deviation (AARD) of 3.65% for a data set including 496 (396 for training and 100 for validation) experimental glass transition temperature. Furthermore, the predictive power of the model is evaluated using the NIST Standard Reference Database #147 data. The model shows an AARD% of 4.3%. (C) 2012 Elsevier Ltd. All rights reserved.