This paper deals with the optimal state observer of nonlinear systems based on a new strategy. Despite development in state prediction for linear systems, state prediction for nonlinear systems continues to be difficult. To acquire a future estimate of the system states, this article first accomplished Taylor series expansion of states in their receding horizons to any given order, and then produced an analytic solution for the prediction error issue, resulting in a closed-form for the nonlinear optimal observer. In proposed observer, the observer gain was optimally chosen among gains obtained from the analytic solution of prediction error problem and satisfied stability condition. Finally, the qualitative simulation results showed the effectiveness of proposed method in the state observation.