This paper presents machine learning technique was used to construct a clinical decision system for the diagnosis of breast cancer. Data gathering, processing, feature extraction, Feed Forward Neural Network (FFNN), a breast cancer diagnostic system, and a breast cancer predicting model were the approaches used. The data gathered contains 23 attriutes of breast cancer. To create the prediction model, the FFNN model was created and trained using the obtained data. Using the MATLAB programming environment, the model was used to construct the best cancer diagnostic system. The tenfold cross-validation approach was used to assess and validate the predictive model. The outcome revealed a regression coefficient (R) of 0.9884 and a mean squared error (MSE) of 1.57e-5, both of which pointed to strong training effectiveness and the capability to accurately predict breast cancer. The outcome of the evaluation of the diagnostic system created with the predictive model was 19.18%.