This paper proposes structured illumination microscopy that utilizes the global information in the curve of modulation depth response to identify overlapping peaks and obtains the height of each surface of the thin film sample analytically with an optimization algorithm under boundary constraints. Measurement of layer thickness distribution and surface morphology reconstruction are thereby achieved in a manner of high thickness resolution, high accuracy, and rapid calculation. According to the simulation analysis, the proposed method improves the thickness detection resolution from 483 nm to 175 nm under ideal conditions. Furthermore, experiments show that the proposed method reduces the number of iterations and offers high repeatability accuracy.