Mining association rules in the database is one of important study in data mining research. Traditional association
rules consist of some redundant information, and need scan database many times and generate lots of candidate item
sets. Aiming at low efficiency in association rules mining using traditional methods, this paper proposes the algorithm
(ISMFP), which is based on intersection for mining the maximum frequent patterns. Firstly, applying the intersection theory
of mathematics, put forwards a number of concepts and definitions. Then gives the process of association rules mining,
and analyzes its performance. After that, the example describes the process of implementation of the algorithm. Finally,
the experimental results show that the algorithm ISMFP is efficient on mining frequent patterns, especially there exists
low threshold of support degree or long patterns.