An advanced adaptive neuro-fuzzy inference system (ANFIS) has been proposed for accurate prediction of the two-phase behavior of CO2/perfluoroalkane mixtures regarding the design of CO2 sequestration and recovery systems. The most important fluoroalkane solvents such as linear (perfluoro-n-octane), cyclic (perfluorodecalin and perfluoromethylcyclohexane), and aromatic compounds (perfluorobenzene and perfluorotoluene) are considered in this study. The proposed ANFIS model can be used for accurate predictions of CO2/perfluoroalkane phase behavior over wide ranges of temperatures and pressures and is a useful tool for the design of environmental friendly CO2 recovery processes. The proposed neuro- fuzzy model outperforms the soft-SAFT (Statistical Associating Fluid Theory) and polar soft-SAFT molecular models both in accuracy and in generality.