We propose a simple visual servoing based on linear approximation of the inverse kinematics. When we use a hand-eye system which has a similar structure as a human being, we can approximate the transformation from a binocular visual space to a joint space of the manipulator as a linear function. This relationship makes it possible to produce the desired joint angles from the image data using a constant linear function instead of the variable nonlinear image jacobian and robot jacobian. This method is robust to the calibration error, because it uses neithor camera angles nor joint angles. We show some experimental results which demonstrates the effectiveness of this method.