Abstract: A control method was proposed, which combines the human-imitating control theory and the optimization capability of chaotic neural networks (CNN). Firstly, a qualitative control law(linear or nonlinear control law)was formed according to human-imitating control. Then, the qualitative control law was quantified by CNN. Hopfield neural network is recognized as a useful tool for optimization problems. However, it is often trapped to a local minimum solution. Therefore, gradually reducing chaotic noise is added to the networks to form a powerful globe optimization algorithm. Applying the proposed control method to a double inverted pendulum, the results of numerical simulations and experiments both demonstrate the valid of this control method.