Due to the complexity of geoscientific data, such as geochemical data, geophysical data and digital remote sensing data, traditional data mining methods, such as cluster analysis and association analysis, have limitations in resources evaluation. In this paper, a clustering algorithm is presented which has the ability to handle clusters of arbitrary shapes, sizes and densities. For association analysis, quantitative association rules aims to deal with the relationships among continuous attributes of geoscientific data objects in resources evaluation. An association analysis algorithm based on the distances between clusters projected on attributes is presented. Applications indicate that the algorithms are effective in real world applications.