The utilization of digital image processing holds significant potential for classifying the ripeness of curly red peppers (Capsicum annuum L.). This study aims to develop an automatic classification method using multiclass Support Vector Machine (SVM) with a linear kernel. Images of peppers, captured using a smartphone camera, were categorized into three classes: ripe, unripe, and semi-ripe. Features such as mean, variance, and range from the RGB color space were extracted for training and testing data. Testing was conducted by dividing the data into training and test sets and employing 10-fold cross-validation. Results demonstrated a classification accuracy of 98.33%. The combination of mean, variance, and range features significantly improved accuracy compared to single features. This research demonstrates the effectiveness of the developed method and its applicability in automated classification systems to support the agricultural sector.