Abstract:With the emergence of a large number of XML data, the field of data mining raises new challenges. Traditional data mining algorithm is oriented relational database and data warehouse, and can not be directly used for data mining in XML documents. From the basic theory of fuzzy sets, by defining the softening properties of the domain partition boundary, a fuzzy association rules oriented XML data mining is proposed, and implement it used the Java language. Experimental results show that the algorithm is correct.