The application of computer vision technologies is hardly new in television sport broadcasts. Most of the developments aim to visually enhance the broadcasts and help the viewers to follow and understand the sport. Recently there have been attempts to automatically analyze sport games. The analysis can be obtained by installing sensors, such as GPS devices on the athletes. There are numerous cases where this is impractical or even against the rules of the sport. In such situations the players’ trajectories can obtained by cameras. In my thesis I consider a particular subclass of these problems. I developed a system to track multiple players and the ball through a football scene. Before I started to design this system, I get acquainted with the similar researches and the applicable computer vision algorithms. Since tracking multiple players through frequent occlusions is not a simple task, I propose, implement and evaluate multiple techniques for each component of the system. The thesis documents the implementation of the system as well as the test that has been performed to evaluate the systems accuracy and precision.