Breast Cancer is one of the most common types of cancer among women. Breast cancer occurs inside the breast cells due to excessive amount increase in production of cells. Most often this can cause death if not cure at a right time. There are many techniques to detect breast cancer and various abnormalities which are described in this report. But, in this research mammography technique is used to deal with the abnormality type: breast masses. These mammograms (X-ray images) of breast masses are stored in the standard mini-MIAS/DDSM databases. To finding the region of interest there are two methods are applied on it these are: segmentation and noise removal by using neural segmentation and thresholding respectively. After the extraction of abnormal part or region of interest, feature extraction is done through using three features: GLCM, GLDM and geometrical feature on which feature selection is applied to get higher accuracy. After calculating the value of each and every feature the classification is done through using method ANN (Artificial neural network) in which 40 mammograms are used to evaluate the terms named as True Positive, True Negative, False Positive, and False Negative with the help of confusion matrix. By using these confusion matrices, the system can understand the stage of each case. Performance evaluation explains that how much effective and beneficial the new research is. Hence, ANN are used to evaluate the performance through defining Accurateness (precision), Sensitivity and Specificity and also compare the results with existing SVM classification technique.