ACC (Adaptive Cruise Control) system has been attracted much attention because it has a potential to reduce traffic accidents and drivers' stress. The system maintains a safe inter-vehicular distance and assists driving by keeping traffic lane automatically. Though the conventional ACC system can observe the preceding vehicles using a millimeter wave radar, it fails in following in case of a curved road because of the narrow detection angle. The present paper develops an ACC system to maintain the safe distance by using a monocular camera even though the tracking course is curved. The inter-vehicular distance and the steering angle is calculated by image processing of the acquired video image. By using the processed data, the proposed ACC system properly controls the vehicle velocity and the steering angle. The effectiveness of the proposed system has been evaluated through experiments using mobile robots instead of real cars. The experimental results reveal that the proposed system surely maintains the inter-vehicular distance to the preceding mobile robot.