Water level predictions are important to the lives of anyone whose livelihood depends on the ocean. Along the Texas Gulf Coast, water level predictive models based on Artificial Neural Networks (ANN) were developed for fourteen locations. These predictions are helpful, but the models could be improved during extreme conditions, such as hurricanes and weather fronts, particularly along the open coast. In an attempt to make more accurate models, wave data from two offshore locations were incorporated into ANNs. The results were then compared to a base model without wave data to determine the importance of the new input. Coastal data including winds and water levels were downloaded from the Texas Coastal Ocean Observation Network website for the Bob Hall Pier station located near Corpus Christi, TX. Wave and meteorological data including wave height and period were downloaded from the National Data Buoy Center website for Buoys 42019 and 42020. The test period was from 2005 to 2010 and included Hurricanes Ike and Dolly. The Matlab computational environment was used for both data processing and ANN modeling. Three ANNs were compared: one for the coastal station only and two other models each including one of the offshore buoys. Each of the ANNs was optimized by changing the number of hidden neurons and delays. Models were compared based on regression and time series response plots as well as the central frequency of 15cm and root mean square error. Results indicate that wave data should be incorporated in future ANN models.