Abstract: Based on the fact that the Elman model of neural network toolbox in MATLAB can approximate any nonlinear continuous function and reflect dynamic features of the system，a method to predict the width spread in the medium plate mill by the Elman neural network was presented. The predictions by the Elman neural network and those by the BP neural network were compared through examples. The results show that the Elman neural network converges more quickly, and its accuracy is higher than the BP neural network.

Key words: metallurgical machinery and automation, MATLAB, Elman neural network, medium plate mill, width spread