In robotics, obstacle avoidance in real-time is a fundamental aspect for path planning of autonomous mobile robots. A dynamic fractional repulsive potential field was recently proposed: the main advantage lies in the consideration of the position, velocity and the nature of danger of each obstacle. One drawback is the lack of continuity of the repulsive force when entering the danger zone around an obstacle. Consequently, acceleration oscillations and energy losses results in a less efficient path planning for obstacle avoidance. Our new proposal of continuous dynamic fractional repulsive potential field aims to correct such a drawback. Simulations illustrates the performances of our method and a comparison with classic potential field methods is provided.