<p>Smart microgrids (MGs) are a potentially effective way to improve the efficiency of energy use and delivery. This research presents a revolutionary real-time economic smart MG operation method that utilizes cutting-edge artificial intelligence (AI) algorithms for dynamic energy management. This research aims to create a dynamic energy management system (EMS) that maximizes the long-term operating expenses of MGs without depending on detailed projections or distribution knowledge of anomalies. Model-predictive control (MPC) and adaptable support vector machine (ASVM) learning are two sophisticated AI approaches that will be used in this study to model the real-time scheduling issue as a finite-horizon Markov decision process (MDP) over a day and develop near-optimal real-time scheduling regulations. The suggested approach entails expressing the problem with real-time planning as an MDP and then using ASVM and MPC learning to derive almost optimum scheduling strategies. The paper compares the MPC technique with an approximation dynamic programming (ADP) framework using a thorough simulation of actual power grid data from the California Independent System Operator (CAISO). To get a nearly optimum policy, the approximation policy iteration (API) approach is used. The success of the suggested strategy in streamlining MG operations is shown by the simulation results. A comparison of MPC with ADP shows that MPC has several benefits, primarily including lower costs. The study considers the effects of AC power flow restrictions and demand response rules, demonstrating the adaptability and resilience of the established technique in a range of circumstances.</p>