Objective With the help of machine learning, the risk prediction model of ischemic stroke in Jiangxi is constructed. Methods With the help of questionnaires, 574 patients with ischemic stroke and 171 healthy people who visited a third-class hospital in Jiangxi Province from January 2020 to December 2020 were obtained. Their basic information and pathological characteristics of ischemic stroke were collected, and the above characteristics were analyzed by machine learning, and the risk prediction model of ischemic stroke was constructed. Results Based on the t test and Mann Whitney test, it was found that there were significant differences in age, symptoms of carotid stenosis or occlusion, and systolic blood pressure between ischemic stroke and healthy people (P<0.05). At the same time, with the aid of AUC evaluation method, based on Naive Bayes model and support vector mechanism, the risk prediction model of ischemic stroke in Jiangxi Province is established for the above indicators, and it is considered that support vector machine performs best (AUC is 0.996 and 1.000 respectively).Conclusion The risk prediction model of ischemic stroke in Jiangxi constructed by this study has high reliability, and there is a strong correlation between ischemic stroke and a variety of pathological characteristics, which should be paid attention to in the followup prevention and intervention of ischemic stroke and precision medical treatment.