Diabetic retinopathy is an ophthalmic inflammation caused by diabetes which ends in visual defacement if not diagnosed earlier and that has two types namely Non-Proliferative Diabetic Retinopathy (NPDR) and Proliferative Diabetic Retinopathy (PDR). NPDR features are present in the earliest stage and systematic detection of these features can improve the diagnosis of the disease severity formerly. Several detection methods exists previously, still there is performance lack on large datasets. The objective of this study is detecting NPDR features from diabetic retinaopathy fundus images of large datasets with good performance level. The study has investigated different fuzzy based systems and to execute the objective; GK_FCM approach is proposed which integrates Gaussian Kernel function in conventional FCM. The execution has four phases, initially the input image undergoes preprocessing using green channel extraction, median filter to enhance the image quality and background removal is performed with extended minima transform technique, mathematical arithmetic operation and pixel replacement method to remove the outlier called Fovea (FV). Further it is segmented for extracting NPDR features such as Microaneurysms (MA), Intraretinal Haemorrhages (IHM) and Hard Exudates (HEXU) using Gaussian kernel with FCM of multiple parameters. Finally the extracted features are visually enhanced on the original input image using post processing operation of multi-class contour tracking (MCT) algorithm comprising different contouring measures. The experiments were done on two online available databases namely DIARETDB0 and DIARETDB1. The performance of the proposed method is evaluated using the validation measures and compared with kernel induced fuzzy algorithms like MKFCM and LKFCM, comparatively the proposed GK_FCM method outperforms. This shows that Gaussian kernel based method can be used for the analysis of the diabetic retinopathy fundus images to detect NPDR features of Diabetic retinopathy. The proposed work has given better results with an accuracy of 98.21%.