Early Prediction done on the right class for a certain disease in the medical field is very critical and the effects of misclassification of the disease could be very risky which may lead to the mistreatment of the patient. The important classification performance measurements in medical fields are error rate, recall, specificity and accuracy. This research aims to focus on these four different measurements and improvement on the misjudgments while classifying a person to a disease that will provide him/her from getting the correct treatment. Thus, the accuracy in classifying such medical data should be at the highest with lowest error. Many attempts are made in order to identify this problem with the objective of high precision and better accuracy with different machine learning algorithms. Where these machine learning help to identify the disease, prediction of the disease risk, decision making with treatment, disease measurement. This research aims to investigate the performance of the different machine learning algorithms. Diabetes dataset with a total 2000 rows and 8 columns have been analyzed. After all the framework and model being tested with ensemble classifiers and profound with the highest accuracy from proposed weighted ensemble method i.e. 97.5 with minimum error rate of 0.025. Which is greater than neural network and majority ensemble technique applied by previous researcher.