For the production of fused magnesia, electrodes currents and electrode circle diameter are two important
parameters, which are high-related with the technical indices. As there are many complex characteristics in the production
process, the setpoints of the two parameters provided by manual operations are often inappropriate, which often can
degrade desired control accuracies. Therefore, the hybrid intelligent setting method is proposed by integrating the casebased
reasoning(CBR), iterative learning control(ILC), PI control, neural network(NN) and rule based-reasoning(RBR). The
applications to the melting magnesia factory in China show the effectiveness of the proposed method.