The description of sorption equilibria of solvents in polymers is a key problem which has to be solved in order to achieve a correct understanding of transport processes through dense membranes. Up to now, this purpose is limited essentially to solvents in elastomers, which are described by Flory-Huggins thermodynamics theory, while sorption of gases in glassy polymers is usually based on a mechanistic one, the so-called dual mode model. Given the difficulties encountered as soon as polymer solvent peculiarities (non-regular mixing enthalpy, network elastic contribution) are taken into account, other types of mixtures can hardly be described by purely thermodynamic models unless complicated expressions are used. A simple mechanistic approach has been developed in order to circumvent these limitations; it is based on the assumption that insertion of a solvent molecule into the polymer solvent matrix will be governed by the intrinsic affinity of the solvent for either a polymer segment or an already sorbed solvent molecule. The model enabling cluster formation description has been named engaged species induced clustering (ENSIC).A series of data related to the most frequently used polymer dense membrane materials have been used in order to check the fitting efficiency of the newly developed expression; it is shown to give a very good description of sorption isotherms of many binary polymer solvent systems, including polar and apolar compounds, in either elastomeric, glassy or thermoplastic membrane materials. The physico-chemical interpretation of the two probabilistic insertion parameters used in the ENSIC model, as well as prospects concerning the extension to multicomponent systems or transport mechanisms simulations are discussed.