Two-parameter cubic equations of state (CEOS) are widely used in process engineering calculations. However, inaccurate liquid density predictions remain a significant deficiency in these equations. To remedy this problem, a volume translation of the CEOS is frequently employed. In a recent work, we presented a volume-translated Peng-Robinson equation of state (VTPR EOS) that is capable of providing accurate density predictions for both saturated- and single-phase regions of pure fluids at high pressures. In the current work, we present an extension of that approach, employing conventional mixing rules, to predict densities of liquid mixtures over large ranges of pressure and temperature. For this purpose, two databases were compiled for vapor-liquid equilibrium and liquid density measurements of 73 binary systems composed of diverse chemical species. The molecular species in the databases ranged widely in terms of molecular size, shape, asymmetry and polarity and, thus, were well suited to test the efficacy of our approach. Overall, the databases contained more than 5000 data points for vapor-liquid equilibrium measurements and over 13,000 data points for liquid density measurements of mixtures. Results indicate that extension of the VTPR EOS to liquid mixtures is capable of providing reliable density predictions for diverse binary mixtures over large ranges of pressure. The VTPR EOS predictions of mixture liquid densities yield errors that are three to five times lower than the corresponding predictions from the untranslated Peng-Robinson equation of state (PR EOS). Specifically, the overall percentage average absolute deviations (%AAD) from the VTPR EOS varied from 1.5 to 3 for binary mixtures. This represents a substantial improvement relative to the untranslated PR EOS, for which errors ranged from 2 to 15%AAD. (C) 2013 Elsevier B.V. All rights reserved.