Abstarct: Breast cancer is the most common cancer and the second leading cause of cancer deaths among women. Currently mammography is the most popular method for detection of breast cancer. Masses and Microcalcificatons are most abnormalities that may appear in breast cancer. Breast abnormalities are presented with a wide range of features. Studies show that 15-30% of breast cancer may be ignored by the radiologists. With the recent advances in the field of digital image processing has occurred, radiologists have the opportunity to reduce the error rate. To help radiologist computer aid detection (CAD) systems have been developed. A recognition system generally consists of preprocessing, feature extraction and classification. In this thesis two methods have been proposed to implement this system and also they have been compared to existing methods. In the both proposed method after preprocessing mammography images, with extract the texture characteristics of the co-occurrence matrix (GLCM), to classify individuals with support vector machine classifiers (SVM) and neural networks adapted. Finally, the second proposed method, the selection of optimal features using particle swarm optimization algorithm (PSO) and support vector machine classifier, the best result in the detection of MC obtain from mammography images.