According to the uncertainty, dynamics and dependency of maize diseases, this paper puts forward a kind of maize disease diagnosis model combined ontology with multi-Agent. The knowledge representation of ontology provides a clear formal specification for the onset periods and symptom characteristics during maize disease occurs. Using the Agent’s intelligence and the division collaboration of Agent modules, we have solved the bottleneck problem of the acquisition and representation of maize disease knowledge and improved the learning ability of the ontology. Experiments show that the model can diagnose maize diseases accurately and efficiently and has a good application prospect.