Breast cancer has an important role in mortality among women. Early and accurate diagnosis is essential for effective treatment. The Computer-Aided Detection (CAD) system can assist healthcare professionals in the diagnostic process. This research introduces a practical approach to enhance the CAD system’s performance in breast cancer diagnosis by analyzing thermal images. The research methodology in the suggested CAD system involves the integration of efficient algorithms during the feature extraction and classification stages, along with a novel feature selection algorithm. The SFTA technique is employed for feature extraction. Then, the EO algorithm is applied to the extracted features. The SVM and DTree classification techniques are then implemented to assess the discriminative efficacy of the selected features, aiming for successful group classification. The suggested method is evaluated using the DMR database. The results demonstrate an accuracy of 98.5%, specificity of 99%, and sensitivity of 98%. The breast cancer diagnostic method that was developed exhibits advantages over other approaches that utilize thermography.