Abstract: In order to solve the problem of class center offset caused by merger of class of traditional hierarchical clustering algorithm, an algorithm of hierarchical clustering is proposed based on correcting class center.The accuracy of proposed algorithm is improved.And it inherits the non-dependnce of the initial clustering center of the traditional hierarchical algorithm.After analysis, the algorithm obtains a good clustering result for the known and unknown number of clusters.It is proved that the proposed algorithm is not only more precise than traditional hierarchical clustering algorithm, but also the clustering results show that the algorithm is effective in image segmentation.