Abstract: A novel control system was designed by means of Hopfield neural network to solve the problem of control systems which are need external parameters. A Hopfield neural network was designed by three neurons, and the self-feedback mechanism to update neurons’ weights was used for the Hopfield neural network to implement the assignment of controlling. The simulation of Hopfield neural network was built by MATLAB platform, and the comparison of neural network output and reference standard output was built for the results. The results showed that  the self-adaption control system based on Hopfield neural network has better precision of controlling. And it can complete common device control and  has better feasibility and convenience.