Purpose/Significance To predict the risk of medication for ischemic stroke patients using a machine learning algorithm based on multi-source data fusion.Method/Process The study is based on the international stroke trial datasets. By fusing features of patient demographics, vital sign examination and medication data, it predicts medication risks using random forest, logistic regression and gradient boosting decision tree(GBDT) algorithms.Result/Conclusion The results show that three algorithms performe well, with the best recall of 91.6% and area under the curve is 0.832 for GBDT algorithm. The machine learning algorithms with multi-source data fusion has good applicability in ischemic stroke medication risk prediction.