The plant electronic medical records formed by the “plant clinic” provide new ideas for the prescription recommendation of crop diseases. How to efficiently mine electronic medical record data and assist crop disease prescription recommendation is still a hot research issue, and needs to be solved urgently at home and abroad. On the basis of summarizing and sorting out the existing domestic and foreign research literature, the key technologies of crop disease diagnosis and prescription recommendation, such as spores recognition based on microscopic image, crop disease diagnosis based on spectrum, crop disease prescription recommendation based on electronic medical records, were systematically analyzed and discussed. The results showed that centering on the infection process of crop disease pathogens, the research on crop disease prescription recommendation based on electronic medical record data mining would become a research focus, guided by intelligent prescription recommendation demand. In the process of crop disease prescription recommendation, due to the characteristics and difficulties of crop disease pathogenesis complex, crop varieties and disease types, disease dynamic changes and characteristics, it would be an important direction to research on the analysis of crop disease pathogenesis, diagnostic reasoning, intelligent prescription recommendation and its application strategy based on electronic medical record data mining. It was of greater practical significance to explore the data mining analysis and research of crop disease electronic medical record based on key technologies such as knowledge graph analysis, big data mining and machine learning algorithm reasoning, and visually analyze the pathogenic mechanism of crop disease, and the correlation between characteristics from the regional macro perspective in order to realize single crop disease prescription recommendation based on diagnostic reasoning, and multiple crop disease prescription recommendation based on semantic matching for practical application scenarios.