A manipulator control method was developed to realize the harvesting of operation for tomato clusters at high speeds. In robot harvesting, both speedy transportation and vibration damping are required when a tomato cluster is transported into a box or a basket by using a manipulator. An input shaping method (ISM) is one of the typical control methods for such control problems. Although the ISM requires accurate natural frequencies of the controlled object, the natural frequencies are different for each tomato cluster. Then, the identification of the natural frequency was combined with the ISM in our method. This identification is based on the data obtained in real time using a machine vision system, a force sensor, and a database on the physical properties of tomato clusters. A numerical simulation study and experiments were conducted and it was verified that the proposed method was applicable to the tomato harvesting robot motion.