Abstract: Rapid and accurate prediction of food shelf life has become a hot issue of concern for consumers, producersand managers. The water activity is a very reliable thermodynamic index to reflect the stability of food and is a key fac-tor to predict the shelf life of food quickly. Based on the analysis and summarization of relevant domestic and foreignliterature, the role and influence of three intervals of water activity≥0.8, 0.5～0.8 and ≤0.5 on food quality are ex-plained based on the mechanism of food quality change and shelf life prediction theory, so as to determine the shelf lifesensitivity index. At the same time, the practical application and shelf life prediction models used for foods with differ-ent water activity are summarized and concluded. Finally, two new technologies for shelf life prediction, hyperspectralimaging detection and bioimpedance-based BP neural network, are introduced to provide references for real-timequality monitoring of food.