The paper uses SQL to preprocess data of Electronic Medical Records(EMR) and divides hypertension patients into 4 groups in terms of sex and the age of 65, namely male <65, female <65, male ≥65 and female ≥65. By use of Apriori algorithm, it studies the association rules between examination results and use of medicine and excavates 29 strong association rules. These provide reference for making plans for clinical diagnosis and treatment of hypertension.