At now, breast cancer ranks second among women in terms of cancer-related deaths, making it a major epidemiological issue. The illness is not caught early enough, and half of the one million women diagnosed with breast cancer annually die from the condition. This research aims to predict the occurrence of breast cancer using various ML algorithms, including Feed forward Neural Network, Random Forest, and Decision Tree, with the goal of reducing the risk of death from this disease, which is a second most common cause of death among women globally. This research uses the Breast Cancer Wisconsin (Diagnostic) dataset to assess ML models that may diagnose breast cancer. The FNN model outperformed RF and DT, achieving the best overall performance with a precision, recall, and accuracy of 97.18%. These results highlight the FNN’s robustness in minimising false positives and maximising true positives, making it a reliable tool for breast cancer diagnosis. To further enhance the accuracy of feature extraction and classification, future research may look at incorporating stronger deep learning models such transformer architectures and Convolution Neural Networks (CNNs). The model’s generalisability and clinical usefulness might be further validated by using bigger and more varied datasets.