Dengue is a mosquito borne endemic disease threatening the world now-a-days. If not treated on proper time, it will lead to death. World Health Organization (WHO) reported, dengue is prevalent in more than 80 countries all over the world. Globally, 50-100 million cases are affected with dengue every year. In order to reduce the mortality rate, there is an urgent need for early diagnosis. This research aims to develop a prediction framework that supports early diagnosis of disease. It summarizes the hidden causes for infection at an early stage, using decision tree classifiers. Decision tree classification is done using clinical dataset from a hospital at south TamilNadu is used for data analysis. The proposed system resulted in prediction accuracy of 86.13%. Number of true positives and true negatives are high. The proposed model identified critical values for symptoms that lead to illness. © 2018 IEEE.