Several algorithms have focused on processing path expression queries. Following those algorithms, all the nodes, matched with path expressions, are participated in computing. In this paper, we propose a novel filter strategy to reduce the number of candidate nodes based on the structure of XML data. All nodes are clustered based on their labels, and path information of each node is kept in bit vectors. Our filter technology mainly depends on high performance of bit operations. The experimental results show that these filter algorithms are effective, scalable and efficient.