The sub-feature classes of geology and mineral resources maps are often closely related to the professional attributes of geological entity, which has an important significance for the refinement and quantify of the geological feature. However there is still a lack of complete division plan. By analyzing the massive geological maps produced in the project of National Mineral Resource Potential Evaluation, and starting from the symbolic expression of geological factors, this article proposed the subclass classification method of geological features based on symbol characteristics. The method can be associated with potential evaluation data model in order to realize dynamic adjustment and optimization of the classification results. Also, the experiment for data of potential evaluation using developed software tools shows that the results are automatic and efficient, which has certain reference value.