In the RoboCup 3D simulation system, a number of important information in the world model,such as the agent coordinate and the soccer ball coordinate information cannot be acquired
directly from the analytical sensor information but through necessary calculations. As the traditional Kalman filter algorithm has some limitations in locating and tracking
soccer, this paper puts forward the moving objects tracking algorithm in the soccer robot system based on Extended
Kalman Filter (EKF). In this project, the moving object considers the soccer ball as an example. The simulation results show that the algorithm can improve the shoot precision of the
robot soccer ball and the accuracy of the goalkeeper robot to save the ball in some degree.