<p>A dynamic management model is developed to improve security and operational resilience in power monitoring networks in isolated microgrid conditions. These systems are subject to frequency and voltage instabilities due to limited generation capacity, low inertia, and large shares of dynamic and pulse loads. The dynamic management model integrates isolation strategies, coordinated load balancing, and security-constrained power management strategies. The Multi-Objective Beetle Antennae Search-driven- Enhanced Pity Beetle Algorithm (MOBAS-EPB) facilitates optimizing multiple objectives concurrently and balancing system stability, operating efficiency, and risk management. The framework includes data preprocessing steps as part of the microgrid security and risk management datasets that handle missing values, Z-score normalization, and dynamically detect security threats, reconfiguration of load sharing to avoid cascading failures, and system integrity. Validation is performed on a modified IEEE 33-bus test system under simulated cyber-attacks and load disturbances using the Power Systems Computer-Aided Design (PSCAD) environment. Experimental results demonstrate that MOBAS-EPB reduces frequency deviations by 41.2%, improves voltage stability by 37.6%, decreases outage duration by 48.5%, and shortens network recovery time by 45.9%. The proposed approach enables adaptive control decisions, effectively managing uncertainties from operational and security factors, and demonstrates the potential of combining intelligent optimization with isolation and load coordination to achieve reliable, secure, and efficient operation of decentralized power monitoring networks.</p>