<p>A microgrid that utilises renewable energy sources is viewed as the most appropriate and cost-effective method to supply electricity. As technology has progressed, energy storage systems have become a viable alternative for stationary power applications, aiding in alleviating the inconsistent characteristics of renewable energy sources. This study investigates the management of energy within a microgrid by incorporating energy storage systems. The innovative hybrid strategy is used in this research merges the Similarity-Navigated Graph Neural Network with Tasmanian Devil Optimisation. The primary aim of this approach is to enhance energy efficiency and improve reliability in the microgrid system. In this framework, Tasmanian Devil Optimisation is employed to fine-tune the parameters of the system while the Similarity-Navigated Graph Neural Network is tasked with predicting these optimised parameters. To implement this proposed methodology, MATLAB software has been utilised, and the current technique is applied to evaluate its performance. The findings indicate that this approach surpasses all existing methods. Notably, the results demonstrate that the proposed method achieves an outstanding efficiency significantly outperforming other contemporary techniques.</p>