MODELS AND CONTROL STRATEGIES FOR VISUAL GUIDANCE OF MOBILE ROBOT IN DYNAMIC ENVIRONMENT ARE STUDIED. THE VISION AND CONTROL PROCESSES FOR ROBOT MOTION IN DYNAMIC ENVIRONMENTS HAVE MANY DISTINGUISHED CHARACTERISTICS THAN THAT IN STATIC ENVIRONMENTS WHERE OBSTACLES DO NOT MOVE. OBJECTIVES OF THE VISUAL GUIDANCE AND CONTROL OF ROBOT MOTION DISCUSSED IN THIS PAPER ARE: FINDING A COLLISION-FREE TRAJECTORY THAT TAKES COUNT OF ALL POSSIBLE MOTIONS OF OBSTACLES IN THE LOCAL ENVIRONMENT; SUCH TRAJECTORY IS MOST CONSISTENT WITH A GLOBAL GOAL OF THE MOTION; THE ROBOT SHOULD MOVE IN A SPEED AS FAST AS POSSIBLE, SUBJECT TO ITS KINEMATIC CONSTRAINTS. TWO APPROACHES FOR THE PREDICTION OF FUTURE MOTION STATES OF OBSTACLES ARE STUDIED: ANTICIPATORY APPROACH, IN WHICH THE MOTION HISTORY OF THE OBSTACLE LINEARLY EXTRAPOLATE THE FUTURE MOTIONS, AND PROBABILISTIC APPROACH IN WHICH THE FUTURE MOTIONS OF OBS