The olfaction visualization detection technique was used to detect the freshness of pork. The images before and after the colorimetric sensor array interacted with pork’s volatile flavors were collected, then image processing techniques was used to get digital signals which reflected the characteristics of pork flavors. Those signals were processed by principle components analysis (PCA), and the former 10 principal components were used as inputs of the BP neural networks to distinguish the freshness of pork. Its discrimination rate was 84.62%. The research results show that olfaction visualization detection technique is useful in the rapid evaluation of pock freshness.