Abstract: The estimation algorithm of pseudo-separate bias and state for nonlinear time-varying stochastic system with zero mean, Gaussian white noise disturbance is extended to the case of the system with nonzero mean and correlated noise disturbance. By using "weakening factor", smoother estimation value curve of the states and the bias can be gotten. Finally, simulation result is presented to verify the effectiveness of the new approach. It shows that, compared with Extended Kalman Filter, the computation amount of the new algorithm is much less, and the stochastic time-varying bias of the system can be estimated exactly.