In this paper, an extended dimension fuzzy adaptive control for nonlinear uncertain stochastic systems with actuator constraints is investigated. The approximation error is introduced into the control scheme as a time-varying function by extended dimension fuzzy logic systems (FLSs). Besides, the sign function as a norm of the FLSs vector, in order to reduce the amount of calculation, will create a challenging chattering problem. Further, actuator failures usually are uncertain in time, pattern, and value and may also occur in practice. In order to handle the above issues, backstepping technology is applied to stabilize stochastic nonlinear system with stochastic disturbance and uncertain actuator failures. It is proved that the asymptotic convergence performance and predetermined transient tracking error performance are guaranteed to be within a pre-established sound whether the actuator fails or not. Finally, simulation results have demonstrated the proposed theoretical results.