In the environment of large data, the application value of data mining and neural network technology in food safety supervision is discussed to provide ideas for the innovation of supervision mode in this field in China. Several important safety supervision models of Chinese traditional food first analyzed. Based on the analysis of the deficiencies of current supervision, the more mature supervision strategy of developed countries such as the United States is used to apply big data related technology to the food safety supervision, so as to make the data information timelier and more open. BP neural network is applied to the analysis of food testing data to predict the risk coefficient of a certain type of food in the subsequent multiple regulatory cycles, and to give early warning.