There are a variety of requirements for future forecasts in relation to optimizing the production of wave energy. Daily forecasts are required to plan maintenance activities and allow power producers to accurately bid on wholesale energy markets, hourly forecasts are needed to warn of impending inclement conditions, possibly placing devices in survival mode, while wave-by-wave forecasts are required to optimize the real-time loading of the device so that maximum power is extracted from the waves over all sea conditions. In addition, related hindcasts over a long time scale may be performed to assess the power production capability of a specific wave site. This paper addresses the full spectrum of the aforementioned wave modeling activities, covering the variety of time scales and detailing modeling methods appropriate to the various time scales, and the causal inputs, where appropriate, which drive these models. Some models are based on a physical description of the system, including bathymetry, for example (e.g., in assessing power production capability), while others simply use measured data to form time series models (e.g., in wave-to-wave forecasting). The paper describes each of the wave forecasting problem domains, details appropriate model structures and how those models are parameterized, and also offers a number of case studies to illustrate each modeling methodology.