In this study, free volume theory (FVT) in combination with perturbed-chain statistical associating fluid theory is implemented for viscosity prediction of petroleum reservoir fluids containing ill-defined components such as cuts and plus fractions. FVT has three adjustable parameters for each component to calculate viscosity. These three parameters for petroleum cuts (especially plus fractions) are not available. In this work, these parameters are determined for different petroleum fractions. A model as a function of molecular weight and specific gravity is developed using 22 real reservoir fluid samples with API grades in the range of 22 to 45. Afterward, the proposed model accuracy in comparison with the accuracy of De la Porte et al. with reference to experimental data is presented. The presented model is used for six real samples in an evaluation step, and the results are compared with available experimental data and the method of De la Porte et al. Finally, the method of Lohrenz et al. and the method of Pedersen et al. as two common industrial methods for viscosity calculation are compared with the proposed approach. The absolute average deviation was 9.7 % for free volume theory method, 15.4 % for Lohrenz et al., and 22.16 for Pedersen et al.
