Diabetic Retinopathy (DR) is the main cause of blindness for diabetic patients. As the exudates are the primary sign of DR, therefore early detection and timely treatment can prevent and delay the risk of vision loss. Automatic screening could facilitate the screening process, reduce inspection time, and increase accuracy, which is vital in ophthalmic treatment, this development of exudates detection will help doctors in detecting symptoms faster. In this research, we use an automatic method to detect exudates from retinal digital images with non-dilated pupils of retinopathy patients; starting by detecting both the optic disc (OD) and retinal vessels, then probable exudates are defined through morphological techniques, in the last main phase, four features are implementing as input data for the fuzzy C-means (FCM) clustering to define the existing exudates in the fundus images. The overall detection performance is evaluated through measuring sensitivity, specificity, positive predictive value (PPV), positive likelihood ratio (PLR) and accuracy. These measures are done by comparing results to hand-drawn ground truth (GT) done by an expert; which are comparatively analyzed. It is found that the proposed method detects exudates successfully with average values of sensitivity, specificity, PPV, PLR and accuracy of 86.3%, 98.4%, 20.8%, 86.2 and 98.4% respectively on the testing studied database.