Categorizing visitors based on their navigation patterns on a website is a key problem in electronic logistics. However, user navigation data and feature vector extracted from it are sparse, and traditional clustering method doesn't solve this problem satisfactorily. As a step forward, a closed repetitive gapped subsequence mining based navigation pattern clustering method is proposed. Feature vector of navigation patterns is constructed with repetitive support of subsequence. A bidirectional projected Euclidean distance based fuzzy dissimilarity is proposed and used as distance measure of feature vectors. Experiment result show that this clustering method is effective and efficient.