The authors hold that the normal transformation of geochemical data is abasis for correct estimation of background and anomalies and clear manifesta-tion of correlation between various elements, that R-mode factor analysis seemsto be an effective method for revealing correlated structures of elements andmaking genetic classification and that the multilevel fitting and interpolationmake up reliable grounds for characterizing spatial distribution pattern ofmineralization factor scores. With such techniques, we performed processing ofdata obtained from 1:50, 000 stream sediment survey (covering520 km²) andsoil survey (12km²) in a certain district. As a result, large quantities of infor-mation were obtained, quite a few anomalies related to mineralizations disco-vered, and the knowledge of ore-controlling conditions deepened, thus pointingout the way for further ore prospecting.