Breast cancer is the most common diseases among women’s world-wide. The survival rate of the women may increase early diagnosis of the disease. Researchers helping the physicians for analyzing and predicting the breast cancer as early as possible using various technologies. This research explores the feature reduction property of Principal Compound Analysis (PCA) on breast cancer decisions support system from wisconsin breast cancer dataset which are analyzed in both two dimensional and 3 dimensional components. The data are reduced to 4 features using chi-square method and evaluated the accuracy of classifiers such as K-Nearest Neighbor (K-NN), Linear Regression(LR), Support Vector Machine (SVM), Random Forest(RF), Decision Tree (DT),Gaussian Naïve Bayes(GNB) and Artificial Neural Network (ANN). This is validated with 10 fold cross-validations. These classifiers are evaluated ,in which the ANN method provides high accuracy of 97.00% and also yields better selectivity and sensitivity rates rather than other machine learning algorithms.