A machine-learning approach to automated building of knowledge bases for soil mapping was presented. Classification tree algorithm was applied to generate knowledge from training data. With this method, building a knowledge base for automated soil mapping is easier than using the conventional knowledge acquisition approach. The knowledge base built by classification tree was used by the knowledge classifier to perform the soil type classification of Longyou area, Zhejiang Province, China using Landsat TM bi-temporal imageries and GIS data. To evaluate the performance of the resultant knowledge bases, the classification result was compared to existing soil map based on field survey. The accuracy assessment and analysis of the resultant soil maps suggest that the knowledge bases built by the machine-learning method was of good quality for mapping distribution model of soil classes over the study area.