The improvement of climate modeling and weather forecasting requires better parameterization of air-sea fluxes. The bulk formulae generally used in numerical models are deduced from fluxes estimated from in situ turbulent measurements, often performed at the top of a mast installed on research vessels. However, the effect of airflow distortion caused by the ship and mast structure has been pointed out as a source of bias on flux measurements and therefore, on their parameterization. Thus, the present work describes an appropriate method to correct turbulent measurements of the wind from airflow distortion. Physical simulations which were conducted in a water channel at the C.N.R.M with a scale model of the ship, provided discrete correction coefficients of the relative wind (speed and direction) varying with the wind azimuth angle and the ship attitude. In order to be able to correct the instantaneous wind vector, whatever stream angle related to the ship, polynomial functions were used to interpolate the discrete experimental coefficients. Then, the measured wind vector was first rotated to restore the mean upstream flow direction and after that, a correction for the wind speed was applied. The dataset which was corrected by this method was gathered during the EQUALANT99 (EQUAtorial atLANTic) experiment which occurred during summer 1999 between Salvador de Bahia (Brasil) and Abidjan (Ivory Coast).