The biases and their sources in monthly turbulent heat fluxes from the Woods Hole Oceanographic Institution (WHOI) analysis, and the National Centers for Environmental Prediction- National Center for Atmospheric Research reanalyses 1 and 2 (NCEP1 and NCEP2) are checked in the climatically representative regions in the tropical Atlantic using the fluxes from the Southampton Oceanographic Centre (SOC) and the pilot research moored array in the tropical Atlantic (PIRATA) as references. For the WHOI analysis, the biases in turbulent heat fluxes mainly exist in equatorial regions which are due to the overestimated sea surface temperature and the underestimated 2 m air humidity. For the NCEP2 reanalysis, the maximum biases, about (40 ± 5) W/m2, exist in southeast and northeast trade wind regions, which are mainly caused by the flux algorithm used because the biases in wind speed and air-sea humidity difference are relatively small. In the equatorial regions, the flux biases in the NCEP2 derived from both flux-related basic variables and algorithm are equally large. Although the estimations of time series trends and air-sea humidity difference of the NCEP1 are improved greatly in the NCEP2, the biases of latent heat flux in the NCEP2 are about 20 W/m2 greater than those from the NCEP1 in the trade wind regions. The result shows that the climatologies and monthly variabilities of the turbulent heat fluxes from the WHOI are more accurate than those from the NCEP1 and NCEP2 in the tropical Atlantic, especially on outside of the equatorial regions.