In the light of the smart grid paradigm, power systems have recently brought about higher levels of adaptability and flexibility. The associated functions in energy management system blocks need to adopt robust strategies to provide greater levels of control and protective service. Under-frequency load shedding (UFLS) schemes can be regarded as a critical function, through which the system integrity and sustainability indices are protected reliably. Conventional UFLS schemes usually suffer from low adaptability and optimal setting in response to changes in the operational characteristics of the power system. This paper seeks to propose an hourly framework for the optimal design of an adaptive UFLS function based on advanced metering infrastructure (AMI) in smart grids. In this portfolio, the target is tailored to fit the main UFLS settings based on the system contingencies and associated operational uncertainties. The simulations are executed on the IEEE 39-bus test system, implemented on a 24-hour scheduling time horizon. The effectiveness of the proposed adaptive model is evaluated with a thorough analysis of the derived numerical results.