Abstract: Clustering is an important field of research for data mining,an improved density-based clustering algorithm was presented based on DBSCAN.This algorithm chooses unlabelled points outside a core object neighborhood but not inside the neighborhood as seeds to expand clusters.Then,according to the different conditions for the corresponding clustering expand so that execution frequency of region queries could be decreased,and consequently the time cost was diminished.Experimental results also show that the time efficiency and the clustering quality of the new algorithm are greatly superior to those of the original DBSCAN.