The Louis scheme and the COARE algorithm (version 3.0) are tested against eddy covariance and inertial dissipation methods for friction velocity estimates in different wind-sea/swell regimes. Atmospheric forcing data, tabulated by Donelan et al. (1997.J Phys Oceanog, 27:2 087～2 099), were collected from a mast on the foredeck ofa SWATH (small water-plane area, twin hull) ship in deep sea off the State of Virginia during the surface wave dynamics experiment. These data are representative of low to moderate wind regimes.The aerodynamic roughness length is determined by using the Charnock relationship. The intercomparison shows that the Louis scheme and the COARE algorithm underestimate the friction velocity by 6% and 3% respectively under pure wind sea conditions, 15% and 13% respectively under cross swell conditions, and 21% and 17% respectively under counter swell conditions. The analysis shows that these underestimations were caused by the method chosen to determine the aerodynamic roughness length because it significantly underestimates the aerodynamic roughness length. It is especially true under the cross swell and counter swell conditions.