Objectives:In allusion to the problem modeling complex fracture structures with crossover and truncation of different faults in the geological environment, we developed a new three- dimensional fault network modeling method by implicit function surface reconstruction.Methods: Firstly, based on the Radial Basis Function (RBF) surface, the three- dimensional fault surface model was implicitly constructed from the fault contour data and attitude information in the sequential geological profiles. Then the binary tree data structure is designed to describe the spatial topological relationship between the different faults and the fault- network model was generated. Finally, considering the uncertainty of fault attitudes information, we defined a probability distribution model for the fault normal vector information, and calculated the information entropy to quantitatively evaluate the quality of the fault- network model.Results: Taking the Shuiyindong gold deposit in Guizhou as the example deposit, we selected the experimental area containing multiple faults and constructed the practical three- dimensional fault network models. The information entropy of the model was calculated and its uncertainty analysis and quantitative evaluation were carried out.Conclusions:The three- dimensional fault- network modeling method can configure the spatial topological structure clearly and construct the 3D fault- network model quickly. The fault- network model is helpful for three- dimensional morphological analysis for the complex fracture structures in the quantitative research.