Current control methods for nonlinear stochastic system have shortcomings such as complex design procedures, high costs, and lack of stability and convergence proof. Considering these problems, based on the equivalent nonlinear system method for obtaining the approximate stationary solutions of the nonlinear stochastic systems, this paper proposes an innovative design procedure for the feedback control of stochastic nonlinear system. The control aims at making the statistical information of the output steady-state probability density function (SPDF) follow those of a target SPDF. Lyapunov function method is presented to prove that the controlled systems do approach to the pre-specified SPDF. Simulation results are given to illustrate the procedure and demonstrate the efficiency.