Breast cancer has been classified as one of the leading causes of mortality among women in different parts of the world. However, it is well established that early identification and precise diagnosis of the illness assures the patient's long survival. The ability to distinguish between benign and malignant tumours at the right time is crucial in the diagnosis of breast cancer. An attempt is made in this work to develop a novel prediction system that explores the full potential of support vector machine (SVM) using a modified grey wolf optimization (GWO) technique. The proposed meta-heuristic algorithm named as Modified Grey Wolf Optimization (MGWO) and Support Vector Machine (SVM) technique (MGWO-SVM) has been developed for extracting valuable information through appropriate selection of relevant features in the popular breast cancer datasets like wisconsin diagnostic breast cancer database (WDBC) for early-stage detection of the disease.