In order to calibrate the SWaN model (North American Regional Reanalysis, version 40.51) for the Caribbean Sea as well as to validate the hindcast conducted at the Center for Oceanographic and Hydrographic Research, CIOH, a wave propagation was performed using, as forcing the wind reanalysis from the North American Regional Reanalysis, NARR. The model calibration was performed through the physical parameterization of whitecapping using the Accumulative Steepness Method. Afterwards, model results were compared with data recorded at the buoys of the Dirección General Marítima between 2006 to 2010 both in time and time and in a probability domains. For the four studied cases (Providence, Turbo, Barranquilla and Puerto Bolivar), the model adequately produced the observed data for the seasonal changes as well as for the extreme events. The lowest linear correlation index was found for Puerto Bolivar´s buoy (0.69). By contrast, the highest correlation index was 0.87 at Turbo's buoy. In general, the validation results show that the model reproduces adequately the actual conditions.