This paper addresses the application of artificial neural networks in control design of DC-DC switching regulators. These regulators are highly nonlinear and time variable. Any small change in operating point as well as changes in line voltage and load causes larger changes in dynamic model, so that the system performance and even stability can be affected. Flexibility of neural networks (NN) during learning process provides a suitable tool for controlling of complex nonlinear systems such as switching regulators. The present paper attempts to show this claim by comparing the performance of closed loop system (using NN-based controller) with the classical design methods. The results have been supplemented by nonlinear simulations.