Abstract: The development of low-inertia renewable energy and the complex interconnection of large power grids have led to new features and forms of modern power systems exhibiting, which are typically characterized by time-varying nonlinearity, uncertainty, data diversity, and local observability, etc. The problems and methods of power system analysis are becoming more complex. As a new technology path of machine learning, the deep learning (DL) has unique advantages in solving complex problems such as power system frequency analysis and control due to its powerful ability of data analysis, prediction, and classification. First, this paper analyzed the basic principle and research progress of DL, the training methods, typical model structures, and application features of DL were introduced. Secondly, the application status of DL in frequency dynamic situation awareness, frequency stability, and safety assessment, frequency control and regulation were summarized, and the adaptability of DL application to each kind of problem was discussed. Then, the application framework of DL in power system frequency analysis and control was constructed. Finally, the development trend of DL and its application in power system frequency were prospected.