Beef was stored at freezing point(-1. 7℃ ± 0. 2℃). Beef qualities( total bacteria count,TVB-N value,red color and off-odour) were determined during the storage and volatile substance of beef was detected using PEN3 electronic nose. The results showed that volatile substance of beef increased during storage and the responses of electronic nose were differences among different storage time. Sensor 2,7 and 9 played a more important role than others in detection and they obtained most of volatile substance information. Beef with different storage time could be distinguished effectively using principle component analysis( PCA) and linear discriminant analysis( LDA). The model of electronic nose could predict well in total bacteria count,TVB-N value,red color and off-odour of beef in at different storage time.