Teff is a crucial crop in Ethiopia, grown extensively in the Amhara and Oromia regions. However, outdated agricultural practices create structural issues. This study addresses the challenge of accurately identifying Teff grain based on theirs geographical growing location through utilizing YOLOv8 algorithms. Previous approaches utilizing digital image processing and machine learning techniques have encountered limitations in segmentation and feature extraction, particularly for small Teff grains. Our proposed model aims to overcome these challenges by applying segmentation techniques, improving system efficiency and scalability, and enhancing overall performance in Teff grain geographical growing areas identification. Through a series of experiments and evaluations, we demonstrate the effectiveness of our model, particularly highlighting the performance of the YOLOv8n-seg model, which achieved an average mAP@50 score of 99.5%. The selection of this model is based on its remarkable performance, low computational complexity, compact model size, and efficient inference time, making it suitable for real-time applications and deployment on portable devices. Our findings suggest promising implications for agricultural monitoring and management, with potential contributions to enhancing food security and informing decision-making in agricultural sectors. Key words:- Deep Learning, Image Identification, Image Processing, YOLO, Teff