This journal paper deals with data-Mining striving as emerging technique which plays the vital role in digging out the significant appropriate information from the vast stream of data collection. The present research focusses on the diagnosis of the brain tumours and the predictions of disease distinguishing the healthy individuals and the patients. To accomplish this predictions, machine learning algorithm Multinomial-Naive-Bayes algorithm in the classification technique to prediction of the results in relevance with the brain tumors disease. The proposed research consists of Collection of dataset, pre-processing technique, Feature-selection method, and organisation of the data in the normalised form, classification implementation and in the generation of the predicted results. These depicted results were subjected to the comparative analysis of the existing previous predictive models with the present proposed work which is superior to them.