In Chaps.  3 and 4 , the actuator faults and actuator nonlinearities have been handled by using sliding mode control (SMC) method. This paper further considers a kind of gain perturbations inevitably occurring in controller, namely, so-called non-fragile problem of controller. Facing with the uncertain Markov jump systems (MJSs), the random characteristics of system parameter uncertainties and controller gain perturbations are described by two independent random variables obeying Bernoulli distribution. A non-fragile SMC scheme is synthesized by co-designing the sliding gains and ideal controller gains to ensure that the mode-dependent sliding surface is reached strictly before the specified time T and the stochastic finite-time boundedness (SFTB) is attained during the whole interval [0, T]. In this scheme, the system parameter uncertainties are suppressed via constructing the controller’s robust term by the probability and bound information. A key feature is that a set of mode-dependent sufficiently small scalars are introduced into some coupled Lyapunov inequalities such that the feasible solutions are easily obtained for the SFTB of the closed-loop systems. Finally, the developed non-fragile SMC approach is verified by a single-link robot arm.