In this paper, clustering methods are presented to enhance the stability of automatic voltage regulators using the efficient adjustment of their respective gains. The results show that implementations of some of the clustering algorithms provide better reliability and stability for the feedback-based voltage regulators as compared to the other methods, namely, a model predictive controller (MPC), a gaussian mixture model (GMM), a self-organizing mapping (SOM) and hierarchical clustering (HC) methods. Specifically, the K-Means clustering approach (KM) provided superior stability but a slower rise time of the output voltage of the voltage regulators as compared to the other methods. Furthermore, coordination of the clustering methods is tested for a 10 machine, 39 bus power grid system. The results show that the clustering approach could be applied to improve the efficiency of voltage regulation methods in smart grids and related cyber-physical systems.