In order to improve the accuracy of crime data mining and effectively combat crime, a data mining algorithm based on the same crime characteristics is proposed. Based on the public security system database, through data preprocessing, a set of data vectors with the same criminal characteristics are obtained for information clustering. The approximate relationship is obtained through the fuzzy set, and the fuzzy set is transformed into a determined objective function according to the Bayesian formula. Based on the objective function, the data set is divided into groups with smaller differences to obtain data mining results. The experimental results show that the algorithm has greatly improved accuracy and reliability in data mining of criminal features.