For an optimum performance of a Wave Energy Converter (WEC) it is very important to adjust the operating conditions of the device with respect to the incident wave characteristics. This means that different wave conditions require different device settings, i.e. geometries, structure dynamics, Power Take Off (PTO), or mooring lines if active mooring lines are deployed. The coupling between the control system of the device and an accurate real time wave prediction tool plays a very important role in the power output of a WEC, especially in the case of floating WECs where the setting of the proper body dynamics is crucial for different wave conditions, i.e. to avoid or enforce resonance, minimize or maximize heave or pitch, etc. In this work artificial intelligence is employed to perform real time wave predictions. The results indicate that if wave conditions in an area close to a wave farm are available (a forecasting horizon of about 10 s is considered), dynamic neural networks can predict the wave conditions in the wave farm deployment area, allowing to set the optimum configuration for each WEC.