In this paper, the authors devise a method that a biped walking robot realizes a learning process of walking stabilization with the cooperation of a human, by considering the learning process of a baby. The authors developed a biped walking robot which has a trunk. They have equipped the robot with a system for measuring the zero moment point (ZMP) of the robot by using universal force-moment sensors and with a mechanical interface for the cooperation. The authors did walking experiments with the biped walking robot, in which a human directly taught the robot the desired ZMP which makes the walking be stable, and the robot learned to stabilize its walking with the cooperation of the human. As a result, the robot obtained the given desired stability by modifying its trunk motion and it achieved stable walking by itself through the process of several learning trials. This kind of function is integral to an anthropomorphic robot, which the authors call a Humanoid, being a human's partner in the future.