One of the most common cancers which is affecting the women worldwide is Breast Cancer. In world, today, Breast cancer has become one of the major problems causing the maximum cancer-causing deaths, and along with this, according to the global statistics, the majority of new cancer cases reported are breast cancer cases. The American Cancer society invests more in the research of Breast cancer than in the research of any other cancer. Early detection of this cancer is very difficult as it does not come with the clear figures, signs, and symptoms. However, only the early diagnosis of BC can help in improving the survival rates of patients as these will be provided with the timely and proper treatment. Further if its detected earlier and properly that whether the cancer is benign or malignant it can prevent the patients to undergo the tough stages of treatment like biopsies, which are very painful. It can be seen here that in order to treat the BC patients properly, much research and work to be done on the classification and prediction of BC cancer types is needed i.e; to find out whether the cancer is malignant or benign. Humans (even the medical professionals) can make mistakes in detecting the cancers, therefore here we require a computer based system which is properly trained and can detect the cancers perfectly with maximum accuracy. Because of its wide use and the most accurate and proper results of analysis, Machine learning has been chosen as a proper tool for classification and developing the predicting models to predict the cancer types. In this paper, we have used the different classification models of machine learning (kNN, decision tree, adaboost, Logistic Regression, Random Forest, bagging) and have provided their results along with the accuracy. We have further developed a hybrid data model- A Voting Ensemble Method. We have tried our best to combine the most compatible algorithms with each other so that it could provide us the more accurate results as possible. This voting ensemble method provided the accuracy of 90%. The maximum and most promising accuracy we came up with, was of Random Forest 99% with very good cross-validation score value. We draw our primary data from WISCONSIN BREAST CANCER DATABASE.