In this paper, the problem of fault-tolerant control for uncertain semi-Markov jump systems with time-varying delays is studied. Suppose the system has two types of failures: actuator failure and actuator stuck, which covers a wider range and is closer to the actual industrial background. The nonlinear function is not considered the Lipschitz condition, which is the most commonly used constraint, but adopts the idea of fuzzy logic to linearize the function. The sliding mode surface based on fault model is designed to obtain the equivalent controller, and then to obtain a closed-loop system. By constructing the mode dependent Lyapunov function, a sufficient condition for the asymptotic stability of the closed-loop system is obtained. Under the above processing method, construct the sliding mode control, which use time-varying multiple adaptive rates to estimate fault, nonlinear function and disturbance, and make the original system reach the sliding mode surface in a finite time. Finally, giving two examples to get the correctness and practicality of the theories.