Abstract: Various phase unwrapping algorithms have been proposed for the last two decades.They mainly can be grouped into two classes：Path-following and global phase unwrapping algorithms.In order to compare their performances,firstly,different type wrap phases were generated according to computer-simulating inclined and Gaussian planes at various noise levels.Then three representative algorithms were chosen from the two classes respectively and their unwrapping phases and root-mean-square errors were calculated.Lastly,experimental data were compared and analyzed theoretically.The results show that minimum cost flow algorithm based on network programming and ZπM algorithm based on Bayesian reference are more robust,effective,and prior to be chosen or improved in practical application.