This thesis explores the use of machine learning models to predict breast cancer survival based on patient demographics and tumor characteristics. The dataset, sourced from the SEER program, includes over 4,000 female patients diagnosed between 2006 and 2010. Six classification models were evaluated: logistic regression, linear discriminant analysis, decision tree, boosted tree, random forest, and neural network. After preprocessing and 5-fold cross-validation, the neural network achieved the highest cross-validated ROC AUC of 0.8636. On the independent test set, it reached an accuracy of 0.9057 and a ROC AUC of 0.8665. These results suggest that neural networks can effectively model survival outcomes and offer potential for use in clinical decision support