In the present study, an equation of state model (i.e., Sanchez–Lacombe equation of state-SL EoS) and an Activity coefficient model (i.e., UNIFAC-FV) are combined with elastic constraints models to assess the effect of polymer crystallinity on solubility of penetrants (e.g., α-olefins, diluents, light gases, etc.) in polyolefins (i.e., binary and ternary mixtures). It is shown that the predictive capabilities of the employed solubility models can be significantly improved when combined with corresponding models accounting for the elastic constraints effects on polymer chains (i.e., Michaels–Hausslein, Banaszak modification of Michaels–Hausslein). More specifically, it is illustrated that for the SL EoS, the fraction of elastically effective chains (f) is the only adjustable parameter which has to be considered to accurately predict the solubility of penetrants in semi-crystalline polyolefins at conditions of industrial importance. In addition, such a combined model results in reduction of the adjustable parameters (i.e., binary interaction parameters, kij), particularly in multi-component mixtures. It is also depicted that the parameter f is a function of polymer crystallinity and temperature in a way that f increases with increasing crystallinity and decreasing temperature. Moreover, it is demonstrated that the incorporation of elastic constraints models in UNIFAC-FV improves its predictive capabilities by about 25%. Finally, a comprehensive sensitivity analysis is carried out to investigate the effect of parameter f on the predictive capabilities of the SL-EoS combined with elastic constraints models.