For the nonlinear multivariable control system, a simple adaptive control calculation method based on neural network was proposed. By using neural network as its constitution unit, the simple implicit adaptive control system is formed. It takes full advantages of the ability of neural network to willfully approach the nonlinear function and its parallel distribution and easy access to hardware application. Moreover, it gives full scope to the property of independence between the simple adaptive control and the controlled object, and its simple structure and flexible variables adjustment permit its adaptability to study. The simulation experiments in the hot rolling looper system achieve desirable control and thus prove its validity and feasibility.