Given the importance of blueberry shelf life in ensuring the quality of blueberries during storage, a blueberry shelf life prediction method was proposed from the perspective of gas sensing information. Three kinds of gas content (oxygen, carbon dioxide and ethylene) and five kinds of physical and chemical indexes for blueberry (rate of corruption, hardness, pH value, soluble solids and weight loss rate) in storage environment of 0℃, 5℃ and 22℃ were monitored, the correlation between gas content and physical and chemical indexes were analyzed, and the shelf life prediction model of blueberry was established by using the BP neural network from the perspective of gas. The results showed that the quality of blueberry was affected by the storage temperature, and there was a clear correlation between the change of gas and blueberry quality. Blueberry shelf life prediction model using the BP neural network had good predictive results. Among them, the prediction error was 0.091~0.191d at 0℃, which was 0.069~0.302d at 5℃ and 0.094~0.338d at 22℃. Predicting the shelf life of blueberries by using gas sensing information had the advantages of simple operation and low cost, which was a useful exploration for the shelf life prediction of blueberries.