The goal of this paper is to propose algorithm to build a car system for moving by label; moving vehicles following the sticker is an idea to reduce effort and time for pushing or pulling and carrying the carts by hand. This idea can be applied to cargo trucks in warehouses or for suitcases that automatically move with the user. The vehicle works in the main steps: Video is continuously recorded and then sent to the server, from that video, the server will detect the sticker in the frame by YOLO algorithm (the sticker has been trained to detect) and determine the relative position of the center of the sticker in the frame, finally sending the order down to the vehicle following that sticker. Because the training data set is small (392 images), the detection accuracy of the object is not high, but under the right conditions it can be determined up to 96%. By increasing the sample set and choosing the appropriate number of steps, it is possible to increase accuracy in worse image conditions.