Cholesterol is a waxy substance that contains fat required to produce hormones and other substances in the body. The excessive cholesterol in the blood vessel can be mixed with other substances and called Low-Density Lipoprotein (LDL). LDL can clog the blood vessel and caused heart disease and stroke. Measuring LDL levels is generally done by taking blood samples (invasive) with the lipid profile test method. This research focused on developing a non-invasive detection system for LDL levels status prediction based on eye image using Convolutional Neural Network (CNN) as a classification model. One indicator of excess LDL levels is a greyish-white ring that surrounds the iris called the corneal arcus. The image processing used the Circular Hough Transform (CHT) algorithm for the localization process and Rubber-Sheet Normalization to normalize the iris region. This LDL level status prediction system used CNN as a classification model with 5-fold cross-validation results in an accuracy of 97.14%.