AbstractThis paper proposes an enhanced three‐layer predictive hierarchical power management framework for secure and economic operation of islanded microgrids. The tertiary control, guaranteeing the microgrid economic operation, is built upon the semi‐definite programming‐based AC optimal power flow model, which periodically sends power references to secondary control. To mitigate uncertainties arising from renewable generations and loads, a centralized linear model predictive control (MPC) controller is proposed and implemented for secondary control. The MPC controller can effectively regulate the microgrid system frequency by closely tracking reference signals from the tertiary controller with low computational complexity. Droop‐based primary controllers are implemented to coordinate with the secondary MPC controller to balance the system in real time. Both synchronous generators (SGs) and solar photovoltaics (PVs) are simulated in the microgrid power management framework. A unified linear input‐state estimator (ULISE) is proposed for SG state variable estimation and control anomaly detection due to compromised cyber‐physical system components, etc. Simulation results demonstrated that SG states can be accurately estimated, while inconsistency in control signals can be effectively detected for an enhanced MPC. Furthermore, comparing with conventional proportional‐integral (PI) control, the proposed hierarchical power management scheme exhibits superior frequency regulation capability whilst maintaining lower system operating costs.