Abstract: A novel electronic nose system was developed for the rapid evaluation of moldy corn. It mainly consists of a thick tin oxide gas sensor array and radial basis function(RBF) neural network. This device can evaluate whether the corn is moldy or not by analyzing the gas emitted from the corn. The detection process was introduced as follows: four feature parameters were picked up from the response curve of each sensor, and then were normalized before being analyzed by principal component analysis(PCA) and RBF neural network. The results produced by PCA were demonstrated that it was hard to distinguish the moldy corn from normal samples, while the accuracy of result prouduced by RBF neural network reaches up over 90%. The novel electronic-nose was proved to be more accurate, more convenient and rapid than the traditional methods.