The high penetration of nonlinear and unstable loads into the grid network severely impacts the quality of power delivered by the grid system. Furthermore, it leads to a disruptive situation of maintaining good quality power in the distribution sectors. Consequently, maintaining a good quality of power is very difficult on the power distribution grid. In this proposed work, the Proton exchange membrane Fuel Cell sustained by a three-phase Series Active Filter (PFC-SAF) is employed to enhance the quality of sensitive load voltage in the distribution grid. The generalized bell-shaped membership function based on the Adaptive Neuro-Fuzzy Inference System (ANFIS) is used to generate the reference voltage signals of the PFC-SAF. The proposed system is modelled and analyzed in a Matlab/Simulink environment. The performance of the PFC-SAF is investigated based on various fault conditions and voltage harmonics. From the obtained results, it has been found that the PFC-SAF effectively regulates the voltage in the distribution grid. Also, the total harmonic distortion of the sensitive load voltage is reduced from 20.86 % to 0.55 %.