A generalized methodology has been developed and successfully applied to determine diffusion coefficient of alkane solvent-CO2-heavy oil systems with consideration of swelling effect. Theoretically, a one-dimensional and one-way mass transfer model incorporating the volume translated Peng-Robinson equation of state (PR EOS) has been developed to describe the mass transfer from alkane solvent-CO2 mixture to heavy oil, which accounts for the oil swelling effect resulted from gas dissolution. The heavy oil sample has been characterized as three pseudocomponents, while the binary interaction parameter (BIP) correlations are tuned with the experimentally measured saturation pressures. Both apparent diffusion coefficients for gas mixtures and individual diffusion coefficient of each component of a mixture are determined once the discrepancy between the measured and calculated dynamic swelling factors of heavy oil has been minimized. The volume translated PR EOS with the three characterized pseudocomponents and the tuned BIP correlations is able to accurately predict the phase behavior of alkane solvent-CO2-heavy oil systems. Compared to the apparent diffusion coefficient, better agreements between the measured and calculated dynamic swelling factors have been obtained by use of the individual diffusion coefficients. Addition of C3H8 into CO2 stream is found to not only diffuse faster into heavy oil than CO2 but also contribute to a larger degree of oil swelling, leading to a faster and enhanced swelling effect of C3H8-CO2 heavy oil system in comparison with the CO2-heavy oil system.