In this work, two non-cubic equations of state, Perturbed Chain Statistical Association Fluid Theory (PC-SAFT) and Simplified Perturbed Hard Chain Theory (SPHCT) were used to predict critical lines and regions for several binary and ternary systems, by fitting one binary interaction parameter (K-ij) for each binary system. Results were compared with experimental critical data, which included inorganic gases, hydrocarbons (light and heavy paraffins, non-saturated olefins and aromatics), alcohols, carbon monoxide and carbon dioxide. These experimental data covered wide ranges of temperatures and pressures which are commonly used in industry. For comparison, the data were also modeled using two cubic equations of state: the well-known Peng-Robinson equation (PR) and the Patel-Teja-Valderrama equation (PTV). PC-SAFIF and SPHCT pure-component parameters were obtained by regression, adjusting pure-component data such as vapor pressure and saturated liquid molar volume (taken from DIPPR 1.2.0). Excellent results were obtained with the PC-SAFT EoS, while the SPHCT, PR and PTV EoS did not have a good performance. The Heidemann and Khalil algorithm, with two nested single-variable iteration loops (an internal loop for volume and an external loop for temperature), was used for locating a single liquid-liquid or vapor-liquid critical point. (c) 2007 Elsevier B.V. All rights reserved.