Identification and classification of brain tumours accurately from magnetic resonance imaging are difficult. The authors adopted a technique using wavelet transform and neural network to identify and classify the tumour. The developed system produced an accuracy of 93.24%.