To ensure the safety and efficiency in high-speed trains’ operation, this paper develops a new multiple-model based fault diagnosis and adjustment scheme for high-speed trains. Such a scheme has capacity of diagnosis for failure modes and values and can track a given speed curve asymptotically. Common faults of traction motors are firstly analyzed; a failure mode set is established; and the parameterized model of each failure mode is obtained. An adaptive estimator is then designed based on the parameterized model of each failure mode to obtain an estimator set, and performance functions based on the estimation error are designed to achieve fault diagnosis for high-speed trains with unknown traction motor failures. An adaptive fault adjustment controller is finally designed according to the information of the diagnosed fault to ensure the stability of the system and the asymptotic tracking of the given speed curve. Simulation results show that the proposed high-speed train adaptive fault diagnosis and adjustment strategy can effectively diagnose and compensate unknown traction faults.