The present invention provides data mining method for protecting an association rule of an original transaction data set. The method comprises: step 1, converting the original transaction data into an authorized data set according to a pre-made interference-added policy; step 2, deriving a support degree, in the original transaction data set, of each item set in the original transaction data set; and step 3, according to the support degree, mining out a frequent item set in the authorized data set so as to mine a data association rule in the original transaction data set. According to the data mining method and system for protecting the association rule of the original transaction data set provided the present invention, by adding an interference item and an interference transaction, not only is an excellent privacy protection effect achieved, but also a really useful data association rule can be mined out, which brings positive effects for commercial application.