In order to obtain a correct phase unwrapping (PhU) result, as well as to alleviate the effect of noise, a phase unwrapping algorithmweighted discrete cosine transform (DCT) algorithm based on phase derivative variance quality weight is proposed. In testifying the validity of the proposed algorithm, the weighted and non-weighted discrete cosine transform algorithms are used to unwrap the simulated and experimental phase maps with artificial random and shot noise, respectively. The two results derived from the different ways are compared with the same unwrapping phase without artificial noise, respectively. It shows that the phase unwrapped by use of weighted discrete cosine transform method is closer to the ideal results. At the same time, the two algorithms are almost equal in computation time. All these prove that in recovering the true phase, our proposed method not only has high efficiency, but also keeps satisfactory noise robustness and reliability.