The present study compares the ability of three semi-predictive approaches, namely, the Global Phase Diagram-based semi-predictive approach (GPDA), the Predictive Soave-Redlich-Kwong (PSRK) and the Linear Combination of the Vidal and Mixing rules (LCVM), to describe the phase equilibrium data in the homologous series methane-n-alkanes. The results obtained for the series under consideration demonstrate, similarly as shown before for the homologue series carbon dioxide-n-alkanes, that GPDA predicts the data more accurately than the G(E)-based models correlate them. In particular, GPDA predicts the critical points very accurately and yields a qualitatively correct picture of the global phase behavior in the series. In contrast, both PSRK and LCVM overestimate the critical pressures and generate false liquid-liquid split in the system methane-n-pentane. Although, the GE-based models yield more accurate results for dew-point data at high temperatures, GPDA predicts these data better at the moderated and low temperatures. In addition, it is clearly better in predicting the bubble-point data. PSRK is more accurate than LCVM in description of the homologues lighter than methane-n-decane, however it fails to predict the data of the heavier ones. (C) 2003 Elsevier Ltd. All rights reserved.