This project customs innovative machine learnings methods to forecast brain strokes to MRI data. Leveraging high- resolution MRI scans, the model seeks to detect subtle changes indicative of stroke risk. By analyzing features like lesion location and intensity variations, it aims to capture the complex factors contributing to stroke occurrence. Additionally, the model aims to classify stroke subtypes, such as ischemic and hemorrhagic strokes, using multi-modal MRI data for personalized treatment strategies. Validation will be performed on diverse patient datasets, assessing sensitivity, specificity, and AUC- ROC. The outcome targets the advancements of a a therapeutically useful instrument for earley stroke prediction, enabling proactive intervention and improved patient outcomes. By integrating MRI and machine learning, this project aims to advance stroke diagnosis and treatment, reducing the burden on healthcare systems globally