In basic unified power quality conditioner (UPQC) configuration, the shunt active power filter and series active power filter are shared with dc-link capacitor. During the time of power quality (PQ) problem, the voltage across the dc-link capacitor is varied from the rated voltage. An adaptive neuro-fuzzy interference system (ANFIS) is one of the recent controllers used for regulating dc-link voltage. In ANFIS controller, the membership function of the fuzzy controller is optimized by using network neuron. But, the weight of the network neuron is changed during the network training process; so the neural network (NN) performance and the control operation are affected. Hence, to improve the control operation, a genetic algorithm (GA) based neuro-fuzzy controller (NFC) is proposed in this paper. In the proposed controller, GA based optimization algorithm is used for optimizing the membership function of fuzzy system. The proposed GANFC based UPQC is implemented and the output performances are evaluated. © EuroJournals Publishing, Inc. 2012.