In this paper, an adaptive genetic algorithm (GA) based neuro-fuzzy controller is proposed for both regulating the DC-link voltage of unified power quality conditioner (UPQC) system and compensating the power quality (PQ) problem. In the proposed adaptive GA, subchromosomes are generated for applying the genetic operation. Then, the sub-chromosome genes are migrated until a best fitness value is achieved. Thus, an optimal solution is obtained more quickly and accurately than the conventional GA. The accurate output of GA is then used for developing a fuzzy interference system. Therefore, the proposed GA-based fuzzy system produces an optimal DC-link regulation voltage. So, the UPQC performance is improved against the PQ problem and an optimal line voltage is injected. Accordingly, the PQ problem is compensated substantially. In the proposed adaptive GA, crossover and mutation operations are carried out with respect to the fitness function. Moreover, the proposed adaptive GA neuro-fuzzy controller based UPQC system is implemented in MATLAB working platform and the output performance is evaluated. © Institution of Engineers Australia 2013.