Scanning acoustic microscopy has shown a potential for detecting a surface layer damaged by machining. In this paper, methods which quantitatively identify the extent of the damage are developed. A V(z) characteristic which is formed as an interference between the acoustic wave reflected from the specimen and that radiated by the surface wave is utilized for the identification. A layer thermally damaged by creep feed grinding was used for this study. First it was successfully demonstrated that the extent of the damage can be graphically generated on a CRT by the respective layers from the surface to the matrix. Secondly, the transmitting velocity of the surface wave was shown to be sensitive to the extent of the damage for a material with small dissipation. Thirdly, the total power of Fourier transformed V(z) was verified to evaluate the damage and to be strongly correlated with micro Vickers hardness for a material with large dissipation.