The paper proposes a dual adaptive control scheme for general class of nonlinear stochastic discrete control systems. The method is a simpler alternative to the well known stochastic adaptive control techniques which is known to be computationally expensive. It is developed under the framework of intelligent control to enhance the performance of the controller for more complex and uncertain plants. The proposed scheme avoids the pre-control neural network training phase by taking into consideration models uncertainty and its effect on tracking, in the on-line control phase. The implementation of this scheme requires finding a suitable method for estimating models uncertainty. For illustration purposes the proposed scheme is applied to linear stochastic systems and the desired results are obtained. A simulation example is also worked out in the paper to demonstrate efficiency of the proposed method.