Attempt is made to define the aerodynamic roughness length with a blowing sand boundary layer. Two methods, the log curve-fit method and power curve-fit method, have been tried to determine the aerodynamic roughness length based on the measured wind profiles in a wind tunnel. The results suggest that a good linear relation exists between the aerodynamic roughness lengths obtained by the two methods. But the power curve-fit method yields greater aerodynamic roughness length for the same blowing sand boundary layer. The aerodynamic roughness tends to increase with sand transport rate. The relationship between the aerodynamic roughness length and shear velocity with a blowing sand cloud is much more complicated than that proposed by Owen and others due to the variation with height of the shear stress caused by the blown sand movement.