New correlations for calculating the simulation parameters of the perturbed-chain version of the statistical associating fluid theory (PC-SAFT) equation of state for nonpolar hydrocarbons are presented in this work. The three PC-SAFT simulation parameters-segment number (m), segment diameter (sigma), and segment segment interaction energy (epsilon/k)-are correlated to molecular weight and density at 20 degrees C for a database of 46 components containing n-alkanes, benzene derivatives, and naphthalene derivatives. The accuracy of the correlations is also tested for cycloalkanes. Contrary to PC-SAFT correlations available in the literature, the correlations developed in this work are optimized to minimize the error in vapor pressure and saturated liquid density for the components in the database rather than matching the respective PC-SAFT parameters. It is found that this approach yields increased accuracy for vapor pressure and density as compared to other correlations in the literature. Additionally, the new correlations are not family-specific, making them useful in modeling mixtures of unknown composition and chemical structure. The developed correlations are tested in modeling density and derivative properties for binary, ternary, and quinary mixtures, each treated as a lumped solvent (or a single pseudocomponent). The investigated mixtures contain n-alkanes, isoalkanes, cycloalkanes, benzene derivatives, and naphthalene derivatives. The calculations require knowledge of only the molecular weight and density at 20 degrees C of the mixture of interest without the need of specifying its composition or chemical structure. The improved accuracy of the proposed approach in modeling volumetric properties of hydrocarbon mixtures makes it an excellent candidate for crude oil modeling applications.