Abstract: Clustering is a promising application area for many fields including data mining, statistical data analysis, pattern recognition, image processing, etc. In this paper, a fast density based clustering algorithm is developed, which considerably speeds up the original DBSCAN algorithm. Unlike DBSCAN, the new DBSCAN uses only a small number of representative objects in a core object’s neighborhood as seeds to expand the cluster so that the execution frequency of region query can be decreased, and consequently the I/O cost is reduced. Experimental results show that the new algorithm is effective and efficient in clustering large scale databases, and it is faster than the original DBSCAN by several times.