This chapter addresses the empirical models of the terrestrial magnetosphere based on large amounts of spacecraft data from many past and present satellite missions. Unlike computer simulations, which derive the expected magnetospheric configurations by solving first-principle equations, the data-based models seek to reconstruct the real-world magnetosphere from direct in situ observations and, in that sense, represent the “ground truth” of the data. The empirical models are based on three cornerstones: multiyear archives of satellite and ground data, a mathematical framework describing the magnetospheric field sources, and parametrization methods relating the model input to the solar wind drivers and/or ground-based geomagnetic indices. Accordingly, this chapter includes three sections devoted to each of the above aspects. In conclusion, we discuss future prospects and challenges, with a particular focus on reconstructing instantaneous real-time magnetospheric configurations, based on simultaneous data of multisatellite constellations of the geospace monitors.