Abstract The recent widespread diffusion of Combined Cycle Power Plants (CCPPs) is due to crucial technological advantages with respect to oil or coal fired plants. Among them, CCPPs guarantee higher efficiency and lower emissions. However, in the context of a deregulated energy market, better performances of the CCPPs control schemes are required during load transitions to allow for their use in grid frequency regulation. In this paper, an innovative Nonlinear Model Predictive Control (NMPC) strategy is applied for the regulation of a detailed nonlinear model of CCPPs, specifically designed for control purposes and validated on real plant data. In particular, the steam temperatures of the Heat Recovery Steam Generator (HRSG) are controlled with NMPC and the plant efficiency is improved at partial load. The obtained results are compared to those achieved with a standard control scheme where the PID regulators have been accurately tuned to optimize the control performance.