Abstract: Abstract: For agricultural machinery equipment, the gap in research and development efficiency and convenience between existing technology and expectation of engineering technical personnel is wide because of its great variety and complex influence factors. With the development and popularization of the three-dimensional (3D) technology, a lot of part libraries, standard part libraries and industry part libraries have been developed in 3D software, but these libraries only achieve simple collection of resource and primitive accumulation of model. However, with the increasing scale of 3D model data, organization management and efficient retrieval of model resource have become the problems needed to be solved. Therefore, in order to realize 3D model's intelligent storage, organization management and efficient retrieval of equipment digital resource database, soybean planting equipment was as the research object; on the basis of pedigree hierarchical division, aiming at the characteristics of soybean planting equipment, the correlation of function and geometry between each part was analyzed and defined by expert evaluation according to function collaboration performance and joining-dismounting performance, and the comprehensive correlation matrix was built by weighted summation method. The unit parts were aggregated by the maximum tree method of fuzzy clustering analysis, so clustering tree was produced. The class number of parts was determined by the threshold. When the threshold was 0.45, parts were divided into 6 classes. Combined with industry standards and function of the equipment practical application, soybean planting equipment was completed and decomposed into category module and function module. And pedigree topology of soybean planting equipment was established, which had the features of systematicness, integrity and program transformation. In pedigree topological map of soybean planting equipment, classification method of planter was the basic composition unit, and the map had the hierarchical structure corresponding to modular decomposition. It was decomposed into a special module, a standard module and several function modules. The same color between modules had more strong correlation, and the number of modules was used to represent cluster value among parts, the smaller value showing the correlation degree of module was higher. Model resources were organized and identified in form of pedigree module encoding according to pedigree topology map basic route and matter-element semantic encoding of "object, characteristics and value", in order to guide the model standardization labeling, and provide references for program recognition. Topological map information with the pedigree semantic features was transformed into program codes by Visual Basic software, which were used for the acquisition and analysis of model information. And fast retrieval of model was realized by the hierarchical matching judgment, and fuzzy retrieval and directory tree browse were the ways for model retrieval. The study results showed that the topological map with pedigree semantic encoding feature could guide organization and retrieval of model resource. Hierarchical fuzzy matching judgment retrieval methods based on hierarchical topology map could not only complete hierarchical retrieval, but also realize two-way interactive information. That was the basis to establish 3D model resources collection. And the study results provide the technology template and basic theory for development and application of similar complex mechanical equipment digital resource database.