A gas sensor array of electronic nose system is used to characteristic values of response curves to detect the freshness change of milk.Stepwise discriminant analysis of SPSS is used to optimize the characteristics

and least squares method and the Bayes discriminant algorithm are used to analyze changes of milk freshness respectively.The results show that stepwise discrimination method can effectively reduce data dimension numbers. Compared with the partial least squares method

the Bayes discriminant algorithm adds data distribution density function to recognize milk freshness

the correctness rate of the recognition is improved.The Bayes algorithm can discriminate the change of freshness of pure milk of goods which are in the storage interval of 12 h.So it is an effective method to analyze the change of milk freshness with an electronic nose and the Bayes discriminant algorithm.