Through the correlation analysis of the main elements of paddy in China in 2015, this paper could provide reference for the data mining of food safety risk monitoring result.Methods The correlation of major elements in paddy were analyzed by Bayesian network, and the elements were grouped according to the results. Canonical correlation analysis between the grouped elements was carried out by R statistics software. Results Bayesian network found that there were direct or indirect correlations among the contents of two groups. The first typical correlation coefficient between the two groups of elements was 0.84 according to the canonical correlation analysis. The negative coefficient of V in the first typical variable was the highest, and the coefficients of Al and TI are high, and negatively correlated with the Al content, while positively correlated with the TI content.Conclusion Bayesian network and canonical correlation analysis could analyze the correlation among the major elements in paddy.