Abstract: With the continuous development of power system and the extensive access of new energy sources, the complexity of the electrical structure and dynamic characteristics of power system is constantly increasing, which makes the frequency safety of the system face more and more severe challenges. Firstly, from the two directions of frequency safety assessment and frequency emergency control, the application of traditional method and artificial intelligence method are analyzed, respectively, and the frequency stability problem of high proportion of renewable energy power system is discussed. The rapid development of artificial intelligence methods provides a new research path for the frequency safety assessment of power systems. Especially, the application of deep learning methods in power systems has significant advantages in solving increasingly complex power system problems. In addition, as the last defense line to prevent frequency instability, frequency emergency control has always been the focus of scholars' research. Reinforcement learning has a strong ability to explore strategies, and how to use it to optimize frequency emergency control strategies is bound to become an important research direction in the future. Finally, based on the characteristics of the new power system, the research prospects of frequency safety assessment and emergency control are proposed.