Brain tumour diagnosis is a complex and challenging part in the medical field. The conventional method of Brain Magnetic Resonance Images (MRI) is of human inspection, which is not an efficient way for large amounts of data. This project presents a method for automatic classification and segmentation of Magnetic Resonance (MRI) brain tumour images using Computer Aided Design (CAD) methods. The use of Neural Networks and Fuzzy Clustering methods has shown high potential in this field. In this project, Probabilistic Neural Network using Radial Basis Function (PNN-RBF) is presented as the classification method. The classification of brain tumour stages is done as Normal, Benign or Malignant. After classification, segmentation of image is done with Spatial Fuzzy Clustering Method. The affected tumour part is extracted from the segmented abnormal images using Morphological filtering. The area of the tumour region is calculated thereafter. The stimulated results show that the proposed system performed efficiently and accurately as the classifier and segmentation algorithm provide better accuracy than the other methods. After segmentation, tumour part is extracted and the area of the tumour part is calculated quite accurately. Various alternative methods for classification and segmentation and their shortcomings are also discussed