In the research field of autonomous robots or vehicle navigation, several works have been carried out in order to allow the avoidance of fixed obstacles. However, the presence of a moving obstacles presents a challenge, particularly when the vehicle moves at high-speed. Indeed, in a robot’s environment, it is not enough to use only the obstacles positions for avoiding them but it is more valuable and necessary to consider their future predicted trajectories. In this paper, we present a method based on both the principle of the Dynamic Window Approach (DWA) which is extended for car-like robot navigation, and Extended Kalman Filter (EKF) which is based on moving obstacle detection and the tracking module. The former are detected and tracked using laser rangefinder and individual EKF for each obstacle. The proposed method is tested in simulation for different scenarios that are close to real environments and has shown satisfactory results.