Abstract: A large number of random discontinuities are widely distributed in rock masses and have significant influences on the mechanical and hydraulic properties of fractured rock masses. In the analysis of the mechanical and hydraulic properties of fractured rock masses, the dominant partitioning of discontinuities of rock masses is an important part, and it is still a key for establishing the three-dimensional (3-D) network model of random discontinuities. A new method is proposed for the dominant partitioning of discontinuities of rock mass based on AGNES. In the proposed method we do not need to determine the centers of every cluster before clustering, and the acnodes or outliers can be eliminated effectively after clustering. Through the comparison of the proposed method and the fuzzy C-means method applied in the artificial and randomly generated data of discontinuities, the following conclusions can be drawn. The proposed method is a better method than the fuzzy C-means method in general cases, and it can get more accurate results by eliminating the acnodes or outliers. Finally, the proposed method is applied to a practical project, and the results are shown to be satisfactory.