Robots and Automation Systems have played important roles in food production and agriculture in recent years. Shortage in agricultural workforce and food contamination due to interaction with disease-infected people have emphasized the importance of robotic implementation in food and agricultural production. This paper introduces a method of using robotic arm UR10e and depth camera OAK-D to classify and sort ripe and unripe tomatoes. YOLOv8 will be used as a detection method to detect four stages of ripeness of the tomato fruit: green, turning, pink, red ripe. The whole system will be implemented with fundamental application of ROS 1 Noetic and Moveit Planning 1 for robotic arm’s motion planning. The RTDE controlling method of UR robots will also be applied to compare the efficiency with MoveIt. A dataset of 452 images of tomatoes (with four different ripeness stages) is prepared for the training process. This project presents a stable method for sorting vegetables and fruit (tomato) using the 6-DOF Robotic Arm system and YOLOv8 Object Detection in order to reduce the labor workforce, increase the quality of fruit and vegetable and also limit human contact to food production line. On the overall, the system has performed well and highly accurately in sorting Red Ripe and Green, Turning, Pink Tomatoes along with the width of each Red Ripe Tomato. The system has illustrated a highly precise model with the overall mAP for 4 labels is 0.924. The results also showed a high value in Precision and Recall. By sampling each type of label for 100 times continuously on the conveyor belt (200 times for Red Ripe), the results are promising when there are 100% Red Ripe tomato being detected, 98% for Green, 74% for Turing and 72% for pink. Finally, the research group propose solution to enhance accuracy of the system.