This paper deals with the synthetic applications of several multivariate statistical methods such as discriminant analysis for two groups, discriminant analysis for several groups, stepwise regression analysis, cluster analysis, factor analysis, and in connection with the investigations of natroreplacement rocks occurring in an iron deposit of Province Hebei it indicates the importance of natroreplacement rocks as a guide for iron deposit exploration. Here the natroreplacement rocks in an investigated region have been divided into four classes, i.e. non-replacement rocks, natroreplacement rocks, intensive natroreplacement rocks and altered rocks of superim-position, with various degree of natroreplacement and mineral constitution. For the purpose of classification of natroreplacement rocks and gradation of natroreplacement intensity, the discriminant analysis both for two groups and for several groups offer satisfactory results. The application of stepwise regression analysis, cluster and factor analyses provides new approach and results for the studies on paragenesis of elements and the genetic relationship between the iron deposit and the natroreplacement. In addition, the authors demonstrate that synthetic application of several multivariate statistical methods in solving geological problems has attracted attention in mathematical geology.