The unified approach and singular perturbation theories are two of the most highly recognized and rapidly developed areas in the control field recently. In this thesis, a stochastic optimal controller design using state feedback is examined for singularly perturbed unified systems with Gaussian noise. The difference between the costs of full order systems and those of reduced order systems is just on the order of epsilon. The robust of a singularly perturbed unified stochastic system is investigated by exploring stability bounds under system uncertainties. A new unified stochastic bound is imported for the investigation. Practical application illustrates the validation of the concepts.