Failures exist ubiquitously in physical actuators, which usually are uncertain in time, pattern, and value. This effect would deteriorate system tracking performance, and even trigger instability. In this paper, we consider the problem of compensating for actuator failures by proposing a fuzzy control scheme for stochastic nonlinear systems. Comparison with existing research on stochastic nonlinear systems, it is a challenge to compensate actuator failures, which is uncertain. Our proposed control scheme is designed on the basis of the adaptive fuzzy logic system and Backstepping techniques for this purpose. It is proven that all the signals remain bounded, the actuator failures is validly compensated. Finally, simulations are provided to illustrate the effectiveness of the obtained theoretical results.