The conventional flow regime map to determine the correlations of interfacial transfer in immiscible multi-phase flow is reconstructed using the fuzzy logic membership function in the present study. Since in the present fuzzy flow regime the interfacial transfer rate is determined by the summation of each term weighted with the its own membership, the present fuzzy flow regime is capable of removing spurious numerical instability in the simulation due to the abrupt change of the flow regime. The fuzzy logic membership functions for various flow regimes have been constructed by the experimental observation in the vertical two-phase flow loop using the impedance signals to measure the void fraction. The impedance signals of the two-phase flow are feed into the feed forward neural network which has the output nodes corresponding flow regimes. The excitement of the output nodes are used to construct the Fuzzy Membership function. Therefore, the present fuzzy flow regime map is not heuristic but phenomenological. It was found that smoother transition of the interfacial transfer terms are produced by the fuzzy logic reasoning based on the membership functions of the both flow regimes near the transition. Therefore, the present fuzzy representation of the flow regime map can be recommended for the computer codes to enhance their stability of the convergence and accuracy.