In the agricultural domain, the variety of data used by organizations is increasing rapidly. Also, there is an increasing demand for accessing these data. Now, the problem of the digital divide causes serious problems in manipulating the distributed information. Based on this condition, this paper presents the intelligent retrieval architecture of distributional agriculture science and technology data which focuses on research of the integration support technology, the concept extending retrial technology based on agricultural ontology and the personalization retrieval technique based on the user model. In the experiment, the intelligent data application platform provided by the paper proves that the architecture is effective.