Joint systems in rock masses are geometrically complex; and investigation of the hydraulic properties is important for understanding the hydraulic behavior of jointed rock masses and stability analysis of rock engineering. A discrete fracture network (DFN) model was designed using C++ language based on VC++6.0 platform to realize a fracture network for seepage analysis. The permeability coefficients

possibility of equivalent permeability tensors of the jointed rock masses were examined by simulating flow through DFN models of varying joint properties. A directionality coefficient was defined to quantitatively assess the principal flow direction in jointed rock masses. The results show that the permeability coefficients are anisotropic according to permeability tensor analysis for rock masses with only one set of joints. The degree of obliquities of joint plane affects obviously the principal direction of permeability. The permeability coefficients increase logarithmically with the increase of joints connectivity rates. In the scenarios where two sets of joints existed

the anisotropic properties vary with the joints geometrical distribution. The permeability anisotropy in a fracture network with two mutually perpendicular sets of joints still exists; and however

the directionality coefficients are examined smaller. For the dip angles whose probability distribution can be effectively by normal function

the permeability coefficients increase with increasing of the standard deviation. The principal direction of permeability runs similarly along the direction of angle bisector between two sets of joints.