More than 75% of U.S. population lives within 100 miles of coastline. The coastal population and infrastructures are increasingly subjected to various coastal hazards (e.g., severe wind, storm surge, coastal inundation, beach erosion, collapse of coastal highway and bridges, and harmful algal bloom, etc.) during hurricanes and extratropical storms. To mitigate the damage caused by these coastal hazards, reliable and timely forecasts of the coastal hazards are needed. This talk presents some work in progress towards the development of an interoperable and integrated regional coastal hazard forecasting system with special focus on the Gulf of Mexico region and the U.S. East Coast. The requirements of such a forecasting system is first presented, followed by a description of the current status of the forecasting system which is based on the integration of a high resolution coastal surge model (CH3D and UnCH3D) and a coastal wave model (SWAN), which are coupled to a basin scale surge model (ADCIRC and UnCH3D) and a basin scale wave model (WaveWatch-III). These models are driven by various forecast wind fields (NAM, GFDL, WNA) provided by NOAA. Examples will be provided to illustrate the performance and scientific robustness of the forecasting system, based on lessons learned during the past three hurricane seasons. Use of grid computing and web service to enable end-to-end regional coastal hazard forecasting will be presented. Several research topics (e.g., ensemble forecasting, hurricane-land interaction, hurricane intensity prediction, improved basin scale model, etc.) aimed to improve the performance, accuracy, and usefulness of the forecasting system will then be discussed.