For the purpose of establishing vertical knowledge graph and auxiliary applications in the agricultural field, a set of agricultural knowledge graph construction methods, calculation frameworks and practical application systems are proposed. Firstly, the existing storage form and knowledge representation of knowledge in the agricultural field are integrated and regularized. On the basis of this data processing, the intelligent construction method of automatic and manual dual mode of knowledge graph in the agricultural field is proposed, and the key technology of entity relationship joint model to extract entity relationship and intelligent retrieval of irregular data. Then, similarity calculation will be used to perform entity knowledge fusion on knowledge graph in the agricultural field, making the graph more standardized, accurate and complete. A good graph is visualized and applied to the mainstream functions of intelligent question answering, which makes the whole system sort out the messy agricultural knowledge and apply it better to better assist learning and research.