Diabetic Retinopathy (DR) is a serious complication of diabetes that affects the eyes and can lead to vision loss and blindness. The current diagnostic process for DR is a time-consuming procedure that first involves a dilated fundus examination by a trained ophthalmologist and then requires the ophthalmologist to examine fundus images manually. This study aims to address this issue by proposing a deep-learning model that automates this task by detecting the severity of Diabetic Retinopathy (DR) grades based on fundus images. The objective is to build a classification model that can classify images into one of the five DR grades of increasing severity. The research is developed in three steps, beginning with pre-processing the fundus images from the APTOS Blindness Detection 2019 dataset. The model architecture is then developed and the pre-processed images are used to train the deep-learning model. The model is then evaluated to determine its effectiveness in predicting DR severity and tuned accordingly. The results show that the proposed deep learning model is comparable to ophthalmologists in grading DR structures, demonstrating the potential for deep learning models to improve DR diagnosis and treatment outcomes.