Abstract: Mining association rules is an important issue in data mining community. Fruitful achievements have been made in this field during the past years. Many algorithms were proposed on the subject. However, effectiveness, scalability of the algorithms and attribute's contribution are still bottle necks to mining association rules. A fast and parallel algorithm for mining association rules based on PC cluster was presented in this paper. The algorithm works on the knots of the PC cluster after reducing and projecting operation according to the excellent properties of the upper triangle matrix. An experiment on the PC cluster testifies that the algorithm has the advantage of improving the effectiveness and scalability of mining association rules.