Inappropriate medication use has become an important factor affecting the safety of rational medication. Most traditional medical anomaly detection systems are based on rules to regulate inappropriate medication use。 In this paper, we model the complex relationships among patients, diseases, and medicine based on medical big data to promote appropriate medication use. More specifically, we first construct the medication knowledge graph based on the historical prescription big data of tertiary hospitals and medical text data. Second, based on the medication knowledge graph, we employ a Gaussian Mixture Model (GMM) to represent patients in groups as physiological features. For diagnostic features, we employ the pre-training word vector BERT to enhance the semantic representation between diagnoses. And to reduce adverse drug interaction caused by combination drug use, we employ a graph convolution network to transform drug interaction information into drug interaction features. Finally, we employ the sequence generation model to learn the complex relationship among patients, diseases, and medicine and provide an appropriate medication evaluation for prescribing by doctors in small hospitals from drug list and medication course of treatment. In this paper, we leverage the MIMIC_III dataset and the dataset of a tertiary hospital in Fujian Province to verify the validity of the model. The results show that our method is more effective than other baseline methods in the accuracy of medication regimen prediction of rational medication. In addition, it has achieved high accuracy in the appropriate medication detection of prescriptions in small hospitals.