Breast cancer is a leading cause of mortality in women all over the world. According to the worldwide cancer statistics, early detection and treatment are keys components for improving the recovery rate of breast cancer and lowering the death rate. Machine learning solutions have been proved to be particularly very successful in exploring the origins of such severe diseases, which requires processing vast amounts of data. In the present study, robust grey wolf optimisation-Random Forest (RGWO-RF) approach was proposed. Our proposed approach based on two steps feature selection process and classification. Modified Grey Wolf Optimizer is used to locate and determine the most significant features. Then, utilizing the prior optimum selections of features, by using Random Forest (RF) classifier to classify breast cancer disease. The reason for using RF it’s robustness and highest accuracy. We apply the proposed approach on Wisconsin Diagnostic Breast Cancer (WDBC) database. The experimental result improve that the hybridation between RGWO for feature selection and RF classifier increase the accuracy rate of classification and demonstrating it’s robustness in identifying the breast cancer.