roses are one of the ornamental plant commodities that are much favored by the public because of their beauty and fragrance. Recognizing the types of roses based on the characteristics of the color an shape of the rose petals is not easy because of the large diversity of colors and shapes possessed by roses, therefore this study aims to classify the types of roses by applying the k-nearest neighbour (K-NN) algorithm baseed on extraction color characteristicsof hue, saturation, value (HSV) an local binary pattern (LBP) texture. The design of the system as an implementation of the classification of the types of roses is by using Matlab. The data used in this study were 120 images with each class taking 24 images. The evaluation results obtained from the application of the k-nearest neighbour algorithm with HSV color feature extraction an LBP texture resulted in the highest accuracy 87,50% with the use of k3 values in the data sharing scenario of 80 training images and 40 test images.