Collective pitch control is the main tool used to regulate the speed and power of wind turbines while operating above their rated wind speeds. The main challenges that face collective pitch control designs are the constraints on the control actions, the unmodeled uncertainties, and the unmeasured system states. A tube based explicit model-predictive output-feedback controller is designed to control the collective pitch angle. The proposed controller is capable of handling the constraints challenge, reducing the on-line computational time and producing the optimal control sequence. Furthermore, the proposed controller is robust against the unmodeled uncertainties. The challenge of unmeasured system states is eliminated since this is an output feedback controller. The performance of the proposed controller is compared to the performance of a gain-scheduled PI controller which is commonly adopted in industry. Simulation results through application to a typical 5-MW offshore wind turbine are obtained. Further, experimental results with hardware in loop are obtained for a reduced scale wind turbine model to demonstrate the feasibility of the controller for real time applications. Simulation and experimental results show the superiority of the proposed controller over the gain-scheduled PI controller. (C) 2018 Elsevier Ltd. All rights reserved.