Forecasting space weather events presents the ultimate challenge to a space physics model. A forecasting model should satisfy not only observational constraints such as the onset time, severity, and duration of actual events but also the practical requirement of timeliness, accuracy, and robustness under realistic conditions. Modern space weather forecasters and users rely on a wide variety of forecast methods, encompassing simple nonlinear regressions, complex empirical (assimilative) algorithms, physical/theoretical models, and hybrid methods. For a thorough understanding of the mechanisms of solar influences on Earth, models must relate remote sensing data and the driving influences of solar events on the magnetosphere/ionosphere in terms of physical mechanisms.