Abstract: The structure determination and parameters estimation for stochastic non-linear systems are presented based on the NARMAX model in, this paper. The orthogonal method is applied which allows the parameters estimation independent of each other in the model. It is based on the "Innovation-Contribution" criterion and some other information criteria that we can select the most significant terms into the model: The orthogonal use of the stepwise-regression algorithm mages the model structure independent of the selected term sequence. Several simulation examples are included to demonstrate the affectiveness of the algorithm.