Abstract: In order to enhance the quality of data stream clustering towards noisy and unbalanced data,an effective twice-clustering algorithm for data streams,TCLUSA for short,was proposed.TCLUSA is based on the simple divide-and-conquer and separability theorems,uses DBSCAN(density-based spatial clustering of applications with noise) to get the average point of each cluster as its local result,and then achieves the final result by clustering all the average points using the k-means.The algorithm keeps all the average points by a layered structure.The theoretical analysis and experimental results demonstrate that the proposed algorithm can enhance clustering quality efficiently when data distribution is abnormal or a high dimensional data stream is dealt with.