I will present the results from the Water In Star-forming regions with Herschel Space Observatory key program (WISH) concerning high-mass protostars. This project is designed to probe with HIFI and PACS instruments the physical and chemical structure of these young stellar objects using water and related molecules, and to follow the water abundance all along their evolution. About 20 sources are targeted covering a wide range of evolutionary stages from cold prestellar cores to Ultra-Compact HII regions. Massive stars are rare but are the main contributors to the matter cycle in the Universe due to their short lifetimes and rapid ejection of enriched material. OB stars dominate the energy budget of star-forming galaxies and are visible at great distances. Their formation, however, is not understood and the classical scheme for low-mass star formation cannot be applied as such to OB stars. The dynamics of such (cluster-forming) regions may either be the monolithic collapse of a turbulent core, or competitive accretion. Water, one of the most abundant molecules in the Universe, might elucidate key episodes in the process of stellar birth and, in particular, may play a major role in the formation of high-mass stars. Our main goals are to use the water molecule as a good tracer of the dynamics of the inner regions, to precise the physical structure of the massive protostars, and to measure the amount of water around it. Using Herschel-HIFI and PACS observations of water lines, and using our modeling of the continuum spectral energy distribution, we analyze the gas dynamics from the line profiles. The spectral modeling tools allow us to estimate outflow and infall velocities, turbulent velocity, and molecular abundances. We present also the water maps of these objects. First results (see Chavarria et al. 2010 and Herpin et al, submitted) indicate that the turbulence is highly supersonic and dominates most water line profiles. In the W43-MM1 HMPO, we measure strong accretion rate, 3.5--4.0 × 10-2 M_&sun;/yr, from the fast infall observed. The estimated accretion luminosity is high enough to overcome the expected radiation pressure. We also find that the turbulent velocity increases with the distance to the center. While not in clear disagreement with the competitive accretion scenario, this behavior is predicted by the turbulent core model.