This paper proposes automatic and error-free classification of brain Magnetic Resonance Imaging MRI for analyzing and understanding human brain. We incorporated four steps for classification process, they are level-3 2D Discrete Wavelet Transform, feature vector normalization, PPCA Probabilistic Principle Component Analysis and last ADBRF AdaBoost Random Forest Classifier. First, we use 2D-DWT is for extracting features of the brain MR image. Then feature vector normalization is used for normalizing the features and PPCA to minimize the dimensionality of the normalized feature matrix. Finally the reduced feature matrix is given as an input to the ADBRF classifier to identify the MR brain as a normal or abnormal images. We applied the proposed scheme for dataset-66DS-66 and dataset-160DS-160. DS-66 consists of 18 normal and 48 abnormal and 20 normal and 140 abnormal in DS-160. We used a 5×5 CV Cross Validation for better performance of the proposed method. Three types of performance metrics, the accuracy, sensitivity and specificity are used for evaluating the performance of our proposed method. © 2018 IEEE.