Abstract: The problem of adaptive faulttolerant control was studied for a class of Markov jump linear systems with partly known transition probabilities. The considered transition probabilities were more general， in which the completely known case， partly unknown with known lower and upper bounds case and completely unknown case were all included. Firstly， the adaptive fault estimator was designed for actuator degradation. Then， based on the values estimated online， the robust controller with compensation effect was designed to guarantee the robust stability of the system with faults. At the same time， the freeconnection weighting matrix method was used to deal with the unknown transition probabilities so that the proposed sufficient conditions were less conservative. Finally， numerical example was given to demonstrate the effectiveness of the proposed method.

Key words: Markov jump system, partly known transition probabilities, actuator faults, adaptive estimation, faulttolerant control