Data mining technology has great application values and potential in the food safety field. The feasibility and advantage of the BP neural network algorithm were explained. The process of data preprocessing was introduced, and the experiment of data mining was designed then realized, focusing on sampling inspection data. Finally, by taking advantage of the mining results, a prediction of food inspection conclusions was put forward which verified the method’s practical value and guiding significance. The experiment indicated that data mining method based on BP neural network has favorable robustness and good accuracy. The predictions of unqualified food’s appearance can lead food safety sampling and inspection work in practice, which can put an end to the occurrence of food safety problems.