Due to the complication of modern industrial process, and the increase of nonlinearity and uncertainty in it, this paper proposed a designing method of fuzzy neural network controller based on rough set. The method combines the Rough set and fuzzy neural network, which can derive control rules from input-output data effectively and rind the minimal set of rules. And this method can solve the "rule explosion" problem. Results obtained from a simulation using MATLAB indicate that the proposed method has good-quality controlling ability and self-adaptive ability when encountering sudden disturbances.