An observer-based adaptive fuzzy backstepping position tracking control method for asynchronous motor stochastic system is designed. The reduced-order observer is constructed to estimate the angular speed. The unknown stochastic nonlinear function in the motor model is approximated by fuzzy logic systems. The problem of "explosion of complexity" in traditional backstepping design is solved by using dynamic surface control technique. Simulation results show that the proposed control method can overcome the influence of stochastic disturbance and ensure the adequately small tracking error.