Two major campaigns (Surface of the Oceans, Fluxes and Interactions with the Atmosphere (SOFIA) and Structure des Echanges Mer-Atmosph¿re, Propri¿t¿s des H¿t¿rog¿n¿it¿s Oc¿aniques: Recherche Exp¿rimentale (SEMAPHORE)) devoted to the study of ocean-atmosphere interaction were conducted in 1992 and 1993, respectively, in the Azores region. Among the various platforms deployed, instrumented aircraft and ship allowed the measurement of the turbulent flux of sensible heat, latent heat, and momentum. From coordinated missions we can evaluate the sea surface fluxes from (1) bulk relations and mean measurements performed aboard the ship in the atmospheric surface layer and (2) turbulence measurements aboard aircraft, which allowed the flux profiles to be estimated through the whole atmospheric boundary layer and therefore to be extrapolated toward the sea surface level. Continuous ship fluxes were calculated with bulk coefficients deduced from inertial-dissipation measurements in the same experiments, whereas aircraft fluxes were calculated with eddy-correlation technique. We present a comparison between these two estimations. Although momentum flux agrees quite well, aircraft estimations of sensible and latent heat flux are lower than those of the ship. This result is surprising, since aircraft momentum flux estimates are often considered as much less accurate than scalar flux estimates. The various sources of errors on the aircraft and ship flux estimates are discussed. For sensible and latent heat flux, random errors on aircraft estimates, as well as variability of ship flux estimates, are lower than the discrepancy between the two platforms, whereas the momentum flux estimates cannot be considered as significantly different. Furthermore, the consequence of the high-pass filtering of the aircraft signals on the flux values is analyzed; it is weak at the lowest altitudes flown and cannot therefore explain the discrepancies between the two platforms but becomes considerable at upper levels in the boundary layer. From arguments linked to the imbalance of the surface energy budget, established during previous campaigns performed over land surfaces with aircraft, we conclude that aircraft heat fluxes are probably also under-estimated over the sea. ¿ 1998 American Geophysical Union