Certain risk factors cause eye diseases. One of the most common eye diseases includes Diabetic Retinopathy (DR), and it can be classified into two stages: early-stage (Non-proliferative stage) and advanced stage (proliferative phase). In this proposed system, eye disease such as Diabetic Retinopathy is focussed. The detection steps involve pre-processing, feature extraction, and classification. The Electro Retinogram Signal (ERG) is recorded for a few seconds, and this captured signal is pre-processed to remove noises. In the processing step, Mel Frequency Cepstral Coefficients (MFCC) features are extracted. During the classifier\'s learning stage, these features train the Support Vector Machine (SVM) classifier. In the testing stage, the patients\' MFCC features from the recorded ERG signal are extracted and compared with the features set used for training the classifier. From this comparison, the classifier will predict and classify DR and its severity. The system\'s average accuracy is 95%, and sensitivity and specificity are 94% and 97%, respectively. It is well-suited for working and aged people since they cannot go for frequent medical check-ups, and clinical diagnosing involves complex procedures.