An inverse analysis method for the thermal response spectroscopy is proposed for quantitative measurement of size and depth of the defects. Sequential thermal response data observed on the surface of objective body under active step heating were processed by lock-in analysis scheme based on the Fourier series expansion. Fourier coefficients synchronizing with sine and cosine waves were calculated, and they are represented in spectroscopic diagram. Fourier coefficients data of certain defect depth and size obtained for various thermal fluctuation periods showed characteristic curves in the spectroscopic diagram. In this study, least residual inverse analysis scheme was applied to the defect parameter determination based on the Fourier coefficient values in the spectroscopic diagram.