This paper considers the output tracking control problem for high-order stochastic time-delay nonlinear systems. By constructing a new Lyapunov–Krasovskii (L–K) functional and modifying the adding a power integrator method, an output tracking controller is well designed such that all signals of the closed-loop system are bounded in probability and the mean square of the tracking error can be adjusted as small as possible. The results are further extended to system with stochastic input-to-state stable (SISS) inverse dynamics. Simulation examples are provided to show the effectiveness of the theory.