This paper describes the dependence of the swell percent (%S) on the soil properties, namely, liquid limit (LL) and cation exchange capacity (CEC). This study provides the statistical model that can be used to obtain an indirect estimation of %S due to CEC and LL. According to the coefficient of correlation (R-2 = 0.91), VAF (91.5%) and RMSE (0.727) indices, multiple regression model has a high prediction performance. Although CEC values play an important role in the swelling behaviour of clayey soils, there is no universally accepted simple quantitative swelling potential classification at present. In this paper, a new swelling potential classification and graph, having four respective zones indicating the low, moderate, high, very high swelling potential, was also developed and proposed. This classification will serve to the engineers in order to evaluate the expansivity of the soils, in practice. (c) 2006 Elsevier B.V. All rights reserved.