The Civil Aviation development of China rely on both the independent research of large transport aircraft and efficient air traffic management system. As the core part of ATC (Air Traffic Control), the fatigue detection and management of ATCs (air traffic controllers) is an important factor affecting aviation safety. In this paper, the recent research progresses on fatigue detection of ATCs are summarized detailly from the aspects of traditional subjective and objective evaluation, their advantages and disadvantages are analyzed in addition. Then, the extracting algorithms of ATCs fatigue feature based on speech analysis are described emphatically. Especially based on the background of artificial intelligence development in recent years, the deep learning model and detection algorithm are proposed to recognize speech fatigue better. Finally, the application prospect of fatigue detection is briefly described.