Optimal controls are typically developed based on low-fidelity models. The predicted performance in terms of energy production might be misleading considering the inaccuracy of the low-fidelity models, especially when control is introduced. Therefore it is necessary to assess the performance of optimal controls in high-fidelity simulations. In this paper, a novel implementation of Sliding Mode Control (SMC) is introduced by defining the sliding surface in an optimal sense (buoy velocity in phase with the excitation force). The performance of the proposed control is simulated in a high-fidelity Computational Fluid Dynamics (CFD) based Numerical Wave Tank (CNWT) simulator by applying ANSYS Fluent. Considering the excitation force information is needed by the control, an excitation force estimator is implemented, which collects the measurements in line with the high-fidelity simulation. This simulation framework provides a convincing assessment of the performance of the controls. Numerical simulation results show the introduced SMC is able to optimize the wave power production with the application of the estimated excitation force subject to significant nonlinearities.