Mixing rules play an important part in the ability of an equation of state to accurately model the behavior of various phases at equilibrium. Many mixing rules have been proposed in the literature. However, due to shortcomings of the mixing rules and/or the equation of state, binary interaction parameters have been introduced to improve the accuracy of models. This work proposes an alternative to binary interaction parameters based on a weighting matrix. It is demonstrated that for the Peng-Robinson EOS applied to prediction of the solubility of various species of industrial importance, such as dyes and pharmaceutical products in supercritical CO 2 , the weighting matrix approach performs slightly better than the binary interaction parameter while providing more flexibility in the model.