The fracture data of ore and rock joints are analyzed on the basis of plentiful interior mines geological survey information. And fractal characteristics of spacing and orientation distribution of rock mass joints are showed by using the fractal theory. In order to find the new methods for rock mass discrimination, an intelligent identification model, which embodies the relations among rock mass stability and uniaxial compressive strength, uniaxial tensile strength, internal friction angle, cohesion force, elastic module, fractal dimension of rock joints spacing, and fractal dimension of occurrence distribution, is established by using the neural network based on chaos optimization algorithm. Fractal dimensions of rock joints spacing and occurrence distribution that embody the all-distributing information of rock mass are showed as follows: The lower the fractal dimension value of joints spacing is, the better the rock mass integrity is. And the lower the fractal dimension value of orientation distribution is, the smaller the joints dispersion degree is. It means that the lower the both of fractal dimension values are, the better the stability of rock mass will be. According to rock mechanics parameters and joints fractal characters, rock mass stability under different geological conditions could be predicted; and a basis for engineering supporting design and construction could be provided by using the intelligent identification model.