The subtle forecast of high-impact weather events by the mesoscale numerical simulation is sensitive to the surface momentum and heat fluxes, so the development of reliable surface turbulent flux parameterization schemes has been a very important research subject. Using the Tropical Ocean-Global Atmosphere Coupled Ocean-Atmosphere Response Experiment (TOGA COARE) data, the authors compared the Blackadar scheme of the Weather Research Forecast (WRF) model against the air－sea turbulent flux parameterization scheme developed recently (i.e. LGLC scheme). The results show that the turbulent fluxes are better estimated with the new scheme than with the Blackadar scheme, especially for the wind stress and the latent heat flux, because the new scheme accurately describes the relationship between the Obukhov length and the Richarson number and takes account of the difference between dynamic roughness and thermal roughness. Meanwhile the LGLC scheme is a non-iterative scheme, and can save more CPU time in comparison with the iterative scheme in the WRF model. This offline test results form a basis for further online tests.