Clustering is one of the most important problems in data mining. Clustering algorithm can classify data without any knowledge about it, and find out the information that valuable. Recently, data mining is more and more widely used in the telecommunication area, but because of some problems, such as the size of the data, the type of the data and the complication of the computation, clustering is not used widely. This article gives a MST algorithm that suit for distribute computing. Combining with the method to represent the similarity that suitable for this algorithm, it designs a new clustering algorithm to solve the problem of sea size data analysis. Then, it shows how the algorithm is realized based on the distribute computing model called mapreduce.