ABSTRACT A Diabetic Retinopathy (DR) is an eye disorder that infects the blood vessels of the retina and sooner or later ends in blindness if now no longer recognized and dealt with in time. The early detection and analysis of DR disorder is vital for retaining patient’s imaginative and prescient. However, the overall performance of current strategies become now no longer enough for early detection of DR. In order to triumph over such barriers in DR disorder analysis, the subsequent tristrategies are proposed. First, Diabetic Fundus Image Recuperation (DFIR) technique is advanced with the goal of classifying the candidate fundus image for diabetic retinopathy disorder analysis. Second, Top-hat Mathematical Transform Fuzzy primarily based totally Feature Clustering (TMTFFC) technique is advanced to obtain correct role of vessel phase in fundus photo and to enhance the overall performance price with human graded orientation standard. Finally, Random Forest Classifier with Prognostic Guidelines (RFC-PG) framework is advanced for early detection of diabetic retinopathy illnesses. The test end result display that the proposed DFIR technique drastically lessens the feature choice time of diabetic retinopathy disorder analysis with the aid of using 38%. Further, TMTF-FC technique improves the clustering performance of disorder analysis with the aid of using 18% with the assist of fuzzy primarily based totally feature clustering. Furthermore, RFC-PG framework improves the sensitivity in early detection of diabetic retinopathy disorder with the aid of using 16% as compared to modern-day works.