One of the most prevalent and deadly illnesses impacting women globally is breast cancer. Although early detection is essential for increasing survival rates, prompt diagnosis is still difficult because of human error in interpretation, expensive expenses, and restricted access to medical services. This work improves breast cancer survival analysis and prediction by utilising machine learning, most especially the Random Forest method. We create a system that can reliably differentiate between benign and malignant tumours by training the model on the Wisconsin Breast Cancer Diagnostic dataset. Through feature selection and hyperparameter adjustment, our study optimises the model and shows that a dataset with only eight essential features may attain almost the same accuracy as a full-featured model.Our Random Forest-based method surpasses previous classification algorithms with an accuracy of over 98%, making it a viable tool to help oncologists make decisions. Our goals are to enhance patient outcomes, decrease diagnostic delays, and contribute to more easily available healthcare solutions by using AI into breast cancer diagnostics. By bridging the gap between technology and medicine, this discovery advances the possibility that early diagnosis can save lives in the future.