Jupiter's magnetosphere contains a gigantic sheet-like structure located near its dipole magnetic equator that contains most of the plasma and energetic particles swirling in Jupiter's magnetosphere. Called the current sheet, it behaves like a rigid structure inside a radial distance of ~50 RJ where the periodic reversals of the Br component are highly predictable. Beyond a radial distance of ~25 RJ, the tilted current sheet lags behind the dipole magnetic equator in proportion to the radial distance of the observer. On the nightside, at radial distances >50 RJ, the current sheet is seen to become parallel to the solar wind flow direction. In this work, we analyze magnetic field observations from all six spacecraft that have explored Jupiter's magnetosphere (Pioneers 10 and 11, Voyagers 1 and 2, Ulysses, and Galileo) to determine the global structure of Jupiter's current sheet. We have assembled a database of 6328 current sheet crossings by using an automated procedure which utilizes reversals in the radial component of the magnetic field to identify current sheet crossings. The assembled database of current sheet crossings spans all local times in Jupiter's magnetosphere under differing solar wind conditions. The new model is based on a further generalization of the hinged-magnetodisc models of Behannon et al. (1981) and Khurana (1992). Four new features of the improved model are that (1) close to Jupiter, the prime meridian of the current sheet (the azimuthal direction in which it attains its highest inclination) is found to be shifted by 2.2¿ from the VIP4 model current sheet (Connerney et al., 1998). (2) In addition to the delay caused by the wave travel time, the location of the current sheet is further delayed because of the sweep-back of the field lines. (3) The signal delay associated with wave propagation is seen to vary both with radial distance and local time, and (4) the current sheet is allowed to become parallel to the solar wind direction at large distances in the magnetotail in agreement with the observations. The new model is much superior at predicting current sheet crossing times than previously published models (especially in the midnight and dusk sectors). The RMS error of fit between the modeled and observed current sheet crossing longitudes is 19.3¿. A comparison of the new model with previous models is presented.