Structured light three-dimensional (3D) shape measurement methods are increasingly used in reverse engineering, aerospace, biomedicine, cultural relics protection and other fields. As a key link in structured light 3D measurement, phase unwrapping plays an important role in accuracy, speed and reliability. This paper reviews the basic principles of phase unwrapping technology, the research status at home and abroad, the advantages and disadvantages of various methods and the future development direction. Firstly, according to the different calculation methods, the existing phase unwrapping methods of structured light 3D shape measurement technology are divided into the following four categories for detailed introduction: temporal phase unwrapping, spatial phase unwrapping, deep learning-based phase unwrapping and other phase unwrapping. Then, the advantages and disadvantages of various technologies are compared in detail. Finally, the characteristics of phase unwrapping technology are summarized and the future research direction of this technology is prospected. Based on the review in this paper, the principles and progress of various phase unwrapping methods can be understood. Moreover, according to the characteristics of different technologies, combined with application requirements and actual measurement conditions, the most effective phase unwrapping method can be selected to achieve accurate 3D shape measurement.Phase unwrapping is a key technique involving multiple application fields, and the phase unwrapping method applied in structured light three-dimensional shape measurement technology is mainly reviewed. According to different calculation methods of phase unwrapping, it can be divided into the following four types: temporal phase unwrapping, spatial phase unwrapping, phase unwrapping based on deep learning and other phase unwrapping. The methods proposed during the development of temporal phase unwrapping can be summarized into four categories: phase unwrapping methods based on gray codes, multi-frequency methods, multi-wavelength methods and phase unwrapping methods based on number theory. Spatial phase unwrapping method mainly introduces the principle and development of the quality guided phase unwrapping and the branch-cut phase unwrapping. The phase unwrapping method based on deep learning mainly analyzes the advantages and disadvantages of various methods in terms of improving measurement efficiency and reducing the number of projection patterns. Other phase unwrapping methods mainly include space-time phase unwrapping, geometric constraints and photometric constraints, etc. The application scenarios and main advantages and disadvantages of the methods are introduced. In order to describe several phase unwrapping methods in more detail, this paper compares and summarizes the performance of several typical phase unwrapping methods. The comparison is made in terms of the number of projected fringe patterns required, measurement speed, noise immunity performance and calculation accuracy. In addition, some conclusions are made to the comparison results. Finally, the future development direction of the phase unwrapping method is summarized, aiming to provide reference for the development and research of fringe projection technology.In this paper, the methods of phase unwrapping are classified and summarized. The temporal phase unwrapping is more suitable for high measurement accuracy and no restrictions on measurement time. Due to the advantage of fast measurement speed, the spatial phase unwrapping method is more suitable for high-speed applications. The phase unwrapping method based on deep learning can solve the shortcomings of the temporal phase unwrapping and spatial phase unwrapping methods to a certain extent. In addition, other phase unwrapping methods are proposed for specific measurement scenarios and measurement requirements. In view of the importance of the phase unwrapping method in the fringe projection 3D shape measurement process and the factors affecting the accuracy, the future development directions include the following five points: reduce the number of projection patterns, enhance noise resistance and robustness, reduce the complexity of calculation, improve the accuracy of phase unwrapping, and realize high-speed real-time measurement.