Vehicle tracking is used by smart transportation systems to see traffic in real time, handle traffic jams, and let cars drive themselves. To see how well the YOLOv8, YOLOv9, and YOLOv10 object recognition models can follow cars in real time, they are put to the test. The UA-DETRAC and BDD100K datasets are used to see how much the models can remember, how quickly they can come to conclusions, and how well they make use of computer power. These models don’t work the same way, as shown by strict statistical significance tests used in the study to show their pros and cons. With the data, we can choose the model that works best for smart transportation in the real world. Viewing how the models work in various types of traffic will help us learn how to use them in real life. Researchers have found better and faster ways to keep track of cars. This study makes smart transportation systems better, roads safer, and people better at using smart mobility apps to make choices. It will be easier for next-generation transportation networks to keep track of cars with the help of the data that researchers and politicians collected.