An object handling apparatus capable of securely holding an object and precisely transferring the held object from a first place to a second place without need of a specific jig for positioning the object on a pallet. A position of an object supplied to the first place is detected by a first visual sensor to obtain a compensation amount for compensating a position displacement of the supplied object. The object is held by a robot hand having position/orientation compensated using the compensation amount and position/orientation of the object held by the robot hand relative to the robot hand is detected by a second visual sensor. A displacement of the position/orientation of the object held by the robot hand from a reference position/orientation is compensated based on the detected position/orientation of the object on the robot hand, so that the object held by the robot hand is precisely moved to have a predetermined position/orientation to be transferred to the second place. The robot hand may be a servo hand having fingers driven by one or more servomotors so that position and force of the fingers in holding the object is controlled.