Abstract: The principle of comprehensive knowledge discovery is proposed in this paper.We first propose some theories and concepts as the base of our research,which include spatial knowledge representation system,spatial object classification,spatial relations,comprehensive knowledge discovery,comprehensive knowledge and spatial union information table(SUIT),etc.In theory,SUIT records all information contained in the studied object.But in reality because of the complex and varieties of spatial relations,we only select some possible factors that we are interested in.The selected factors constitute the data to be processed.In this study,we select spatial association as the research emphasis,which was defined as sharing voronoi edge between spatial entities by C.M.Gold in 1992.The index value of spatial association is also introduced in our study.In order to find out the spatial association for spatial entities in a given study area,a highly efficient comprehensive knowledge discovery algorithm called recycled algorithm(RA) is also suggested.As an example,a case study about comprehensive knowledge discovery for spatial association and attributes is studied.We compare the yields of different agriculture crops in two areas with similar climate characters and find that the difference of the same crop is notable.Through comprehensive knowledge discovery,we find that the reason lies in the spatial association relations.The study shows that the principle and methods proposed in this paper have an effective impact on comprehensive knowledge discovery and the discovered knowledge is both valuable and understandable.