A white-light interference algorithm based on spatial frequency-domain analysis is described, which permits a high-resolution surface measurement by removing 2π ambiguity contained in phase information. In the frequency domain analysis, we can extract the coherence and phase profiles simultaneously, where the coherence profile is free of 2π ambiguity but contains the measurement errors caused by background noises, while the phase profile is precisely the opposite. By combining the coherence and phase information, we remove the 2π ambiguity contained in phase information successfully. In addition, a differential analysis method between adjacent pixels is adopted to deal with the local 2π phase-jump errors, which are caused by the background noises and the light source fluctuations. This technique is highly efficient since it does not require enormous computation. Both theories and experiments are elaborated in detail to verify the validity and stability of this method.