The perturbed-chain statistical associating fluid theory (PC-SAFT) was used to predict the solvent swelling extent of vitrinite- and inertinite-rich coals. These two maceral different coals yet similar in rank and carbon content were modeled as a mixture of pseudo-components according to molecular-weight fractions (obtained from laser desorption ionization time-of-flight mass spectrometry). The pure-component parameters for the solvents used for solvent swelling (pyridine, N-methylpyrrolidone, and the binary mixture of carbon disulfide and N-methylpyrrolidone) were determined by simultaneous fitting to experimental vapor pressure and liquid density data. The predicted swelling trends obtained from the PC-SAFT were comparable to experimental swelling extents. This approach may be a promising tool for solvent-coal interaction predictions.