Breast cancer is the second leading cause of cancer deaths in women worldwide and occurs in nearly one out of eight women. Currently there are three techniques to diagnose breast cancer: mammography, FNA (Fine Needle Aspirate) and surgical biopsy. Generally, breast cancer is a malignant tumor that starts from cells of the breast. A malignant tumor is a group of cancer cells that may invade surrounding tissues or spread to distant areas of the body. Classification of cancerous cells as either malignant or benign is a serious and life-altering decision. If the physician diagnoses a benign cancer as malignant, the patient may undergo unnecessary testing and damaging to the patients health and of a physician diagnoses a malignant cancer as benign, the consequences can be much worse. These consequences being rampant spreading of the cancer resulting in extreme sickness or death. Breast Cancer Prediction System is an automatic diagnosis system for detecting and classifying breast cancer based on feed-forward back propagation algorithm. The model to classify breast cancer tumor based on symptoms that cause the breast cancer disease is proposed. Several neural networks model were created and trained using different number of neurons in hidden layer. An experiment using back propagation algorithm is done to investigate the potential of neural network model as outcome classifier. From the experiment, it is found that neural network model that has hidden layer 7 achieve the highest accuracy among others. The proposed feed-forward backpropagation algorithm performance is compared with other 2 classifiers and it showed the feed-forward backpropagation algorithm is the best classifier to predict breast cancer disease. The system provides more cost-effective and easy-to-use systems for supporting clinicians. For the breast cancer tumor diagnosis problem, experimental results show that the concise models extracted from the network achieve high accuracy rate of on the training data set and on the test data set. Breast cancer tumor database used for this purpose is from the University of Wisconsin (USI) Machine Learning Respiratory.