This paper presents a real hybrid power system in Egypt, which includes both conventional generation units and Renewable Energy Sources (RESs) for studying the Load Frequency Control (LFC) problem. The conventional generation system in the Egyptian Power System (EPS) is decomposed into three dynamic subsystems; non-reheat, reheat and hydro power plants. Moreover, real wind speed data extracted from Zafarana location in Egypt is used for achieving a realistic wind power study. Each subsystem of the EPS has its own characteristics compared to the others. Moreover, the physical constraints of the governors and turbines such as Generation Rate Constraints (GRCs) of power plants and speed governor dead band (i.e., backlash) are taken into consideration. Therefore, this paper proposes a decentralized controller for each subsystem independently, to guarantee the stability of the overall closed-loop system. Hence, an optimal PID controller-based Particle Swarm Optimization (PSO) algorithm is proposed for every subsystem separately to regulate the frequency and track the load demands of the EPS. The performance of the proposed decentralized controller of each subsystem is compared to the centralized one under different operational scenarios. The EPS is tested using the nonlinear simulation by Matlab/SIMULINK. The obtained results reveal the superior robustness of the proposed decentralized controller against different load disturbance patterns, real wind power fluctuations and EPS uncertainties.