This paper presents a model predictive controller (MPC) operating a solid oxide fuel cell (SOFC) gas turbine hybrid plant at end-of-life performance condition. Its performance was assessed with experimental tests showing a comparison with a proportional integral derivative (PID) control system. The hybrid system (HS) operates in grid-connected mode, i.e., at variable speed condition of the turbine. The control system faces a multivariable constrained problem, as it must operate the plant into safety conditions while pursuing its objectives. The goal is to test whether a linearized controller design for normal operating condition is able to govern a system which is affected by strong performance degradation. The control performance was demonstrated in a cyber-physical emulator test rig designed for experimental analyses on such HSs. This laboratory facility is based on the coupling of a 100 kW recuperated microturbine with a fuel cell emulation system based on vessels for both anodic and cathodic sides. The components not physically present in the rig were studied with a real-time model running in parallel with the plant. Model output values were used as set-point data for obtaining in the rig (in real-time mode) the effect of the fuel cell system. The result comparison of the MPC tool against a PID control system was carried out considering several plant properties and the related constraints. Both systems succeeded in managing the plant, still the MPC performed better in terms of smoothing temperature gradient and peaks.