With the aging population and labor shortage, the proportion of labor costs in the production costs of tomato harvesting is increasing, making the research and development of tomato harvesting robots urgent. This paper develops a cherry tomato harvesting robot, consisting of a mobile chassis, lifting platform, robotic arm, end effector, depth camera, and control system, to achieve automated walking navigation, harvesting, and basket loading. This paper outlines the system design and harvesting scenario of the tomato harvesting robot, details the design and implementation of the robot’s chassis and walking route planning, end effector design, tomato recognition and positioning, robotic arm motion planning, and control system. Furthermore, the architecture and control flow of the software system are elaborated. Finally, field tests validate the robot’s performance and harvesting success rate. Tests show that the overall accuracy and recall rates of the visual recognition model for tomato bunches are 85.04% and 88.71%, respectively, and for the stalks are 82.72% and 81.51%, respectively, with a harvesting success rate of 70.77%. These efforts provide beneficial exploration for the future commercial application of cherry tomato harvesting robots.