This study examines the sensitivity of a simulation of the tropical Atlantic circulation to the choice of climatological wind stress forcing. Three climatological wind stress data sets, representing different classes of data products, are used to drive a high-resolution general circulation model of the Atlantic Ocean. One set is from historical surface marine datasets compiled by Hellerman and Rosenstein. The second climatology is based on surface wind analyses from the European Centre for Medium Range Weather Forecasts. The last climatology is derived from an integration of the National Center for Atmospheric Research Community Climate Model, version 2, at T42 resolution (CCM2). The analysis of the model sensitivity to these different wind stress climatologies focuses on the mean seasonal cycle and low order statistics of the simulated upper ocean circulation in the tropics and its comparison with observations. In the North Equatorial Countercurrent (NECC) the three simulations agree with each other and with observations reasonably well. The differences that do exist in the simulation of the NECC are primarily in smaller-meridional-scale features and in the phase of the onset and decay. There are much larger differences among the simulations near the equator. Most of these differences in the equatorial circulation are attributable to differences in the cross-equatorial meridional wind stress. The ECMWF climatology has relatively weak meridional stress and correspondingly weak equatorial upwelling. The CCM2 climatology has unrealistically large meridional stress and equatorial upwelling. ¿ American Geophysical Union 1995