The authors tried to extract alteration anomalies in spectral characteristic space (scatter plot) in view of the limitation of the traditional methods. The scatter plot takes on an anisotropic feature in associated distribution of RS data’s grey scale. The distribution is usually combined by oval clusters. Parameters of oval clusters are acquired sequentially by applying U-Statistic method in the frequency of the scatter plot. Through mapping the points inside the oval into RS image and interpreting visually, the spatial distribution of alteration anomalies is obtained eventually. In this paper, this new method was described with the instance of Bayinshan area in Qinghai province. Other data acquired were also comparatively studied, and it is found that the anomalies of ferruginization and argillation are consistent well with each other. This new method has a better performance than PCA in the study area.