Apples are one type of fruit that is increasingly popular in Indonesia. This fruit is not only popular because it tastes good, but is also rich in nutrients and fiber which are beneficial for the health of the body. Along with the development of the agricultural industry in Indonesia, domestic apple production is also increasing. This study aims to classify types of apples based on RGB color using research methods that include apple image data collection, RGB feature extraction, data division with k-fold cross validation, classification model with Naive Bayes. This method utilizes color features taken from apple images as input to determine the appropriate class or type of apple. The test results show that the accuracy for the sweet level has a value of 100%, for the medium level it has a value of 86.66% and for sour it has a value of 80%. The average accuracy of the Naïve Bayes method is 88.88%. Classification results using the Naïve Bayes algorithm.