Abstract: Taking the groundwater system in mining region as example,the applied research to classify and identify the groundwater occurred in different subsystems and stopes was actualized on the basis of hydrochemical data analysis with the techniques of multivariate statistics.Taking the concentration of the major ions dissolved in groundwater as variables,the factor analysis was performed to reduce the dimensions of variables and identify the underlying variables or the orthogonal factors which reflect the hydrochemical characteristics of groundwater and can be then used in cluster analysis.As the results of hierarchical cluster analyses,a set of statistically derived typical samples were obtained,which hydrochemically represent the groundwater in individual subsystems.Multigroup Bayes discriminant analysis stepwise method was conducted with the subsystems as groups and the major initial ions as variables,and then a set of linear discriminant functions was obtained.The functions were applied to predict the sources of some random water samples on the basis of the variables with unknown group membership and proved to have the ability of differentiation with statistical tests.The results show that it is an effective and practical way in classifying and identifying of groundwater or gushing water sources with high stability.