Abstract: With the flourishing development of high-speed train in China, research on the fault prognostics and health management has been paid more and more attention during its service and maintenance. Axle-box bearing is a key rotating component of the bogie in high-speed trains. Under the complex wheel-rail interaction, it is prone to cause in-service failures due to fatigue, overload, etc., which seriously affect the operational efficiency and safety of the train. However, the existing diagnosis methods cannot meet the dynamic and systematic requirements of modern high-speed train. It is necessary to develop more effective health monitoring and diagnosis techniques. In this paper, the state-of-art condition monitoring and fault detection systems and methods for axle-box bearing in maintenance, wayside and on-board are firstly introduced. Then, a review is presented including theoretical modeling and dynamic analysis of the coupled system of axle-box bearing and high-speed train, fault diagnosis of axle-box bearing based on signal processing and machine learning technologies from both positive and negative perspectives of dynamics. At last, the crucial points for further exploring methods and technologies in prognostics and health management of the axle-box bearing are summarized.