Breast Cancer is one of the most common illnesses in recent years. Diagnosing cancer at early stages can have a considerable effect on the therapy, so that many several attempts have been made for diagnosing this illness at its first stage recently. Mammography imaging is the most commonly used technique to detect breast cancer before appearing the clinical symptoms. Extracting features which facilitate cancer symptoms detection without significant decrease in sensitivity, minimizes false positives and is of great importance. Microcalcification is an important indicator of cancer. In this research a new method for detecting microcalcifications in mammography is presented. Due to the ability of wavelet transform in image decomposition and detaching details, it can be used to expose this symptom in mammograms. In this work, a two dimensional wavelet transform is performed for feature extraction; and these features are used to diagnose cancer symptoms in mammography images. After the feature extraction step, classification is done using Support Vector Machine (SVM). In the performed evaluation, Regions of Interest (ROIs) with different dimensions have been used as input data and the results show that the proposed feature extraction method can have a significant impact in improving the performance of detection systems.