Abstract: k-means Clustering Algorithm is widely used and is sensitive to k. The performance of the k-means Clustering Algorithm primary depends on the optimization of k. In this paper, the progresses on k-means algorithm were summarized. Firstly, some representative k-means algorithms about the optimization of k were outlined. Secondly, several known data sets were selected to test some typical k-means algorithms. The work will be valuable for data clustering and data mining.