Proper classification of harvests is crucial to maximize crop yield, ensure high-quality exports, and optimize storage conditions. This study presents the development of a real-time image processing algorithm to detect the maturity stage of scotch bonnet peppers (Capsicum chinense) using Convolutional Neural Networks (CNNs). The algorithm classifies the peppers into three stages: unripe, moderately ripe, and ripe. It demonstrates 89.04% testing accuracy and 91.6% training accuracy, showcasing its effectiveness in real-time maturity detection. This cost-effective and intelligent solution has significant implications for the agricultural industry, potentially reducing postharvest losses and optimizing export quality.