A trajectory generation method based on an artificial potential field approach may be useful for real-time motion planning of a moving robot, since it is effective and computationally much less expensive than other methods using global information on a task space. In the present paper, we argue that dynamic behavior of a generated trajectory and control of the direction of motion of the moving robot shold be taken into accout within the framework of the artificial potential field approach for practical use, and propose a new method using active deformation of the artificial potential field. The method can regulate movement time from an initial position to a goal, and velocity profile of the trajectory as well as the position and the direction of motion of the robot by changing the orientation and the shape of the artificial potential ellipse.