Phase unwrapping is an essential problem in phase measuring profilometry. In this paper, phase unwrapping is modeled as a congruence problem. Based on robust Chinese remainder theorem, the frequency selection rule is proved by analyzing the uncertainty of wrapped phase because of the intensity noise. Then, this rule is adopted to design the frequency unwrapping algorithm and the parameters are determined with high confidence probability. After that, the closed-form algorithm of robust Chinese remainder theorem is employed to determine the correspondence. Finally, the quantitative comparisons are implemented in our developed structured light measurement system. The experimental results show that the frequency selection rule is valid, and robust Chinese remainder theorem is less sensitive to the phase noise than the traditional Chinese remainder theorem. If the frequency selection rule is satisfied, the accuracy of the unwrapped phase can achieve nearly 100%, otherwise, the performance is deteriorated.