This article presents a recent study carried out to evaluate the effect of fluid characterization on liquid drop-out predictions for constant volume depletion (CVD) of gas condensate fluids using the Peng-Robinson equations of state (EOS) model. The work is based on 11 different gas condensate fluids of widely varying overall fluid compositions and C7+ plus fraction properties. The accompanying experimental data on dew point pressures and liquid drop-out from CVD tests are employed to evaluate the effect of fluid characterization on predictions. Initially, all predictions were based on uncharacterized C7+ fraction. Subsequently, all C7+ fractions were characterized using the methods proposed by Katz and Pedersen, respectively, and predictions were carried out on the basis of characterized fluids. All predicted values reported in this work are purely predictive, i.e., C7+ data, and the characterized fluid data has been used as it is, without tuning the EOS models. The performance of the uncharacterized and characterized fluids is compared. The study demonstrates that EOS predictions are significantly improved when the C7+ fraction is described by several pseudo-components and an extended higher plus fraction rather than treating it as a single lumped component.