Path Planning is critical for a mobile robot (MR) to navigate safely towards its target position, without collision with the surrounding obstacles. There are many aspects that need to be considered in path planning; computation time, path smoothness, and path length. The Artificial Potential Field (APF) approach is popular for solving the path planning problem. It is a unique approach that has a simple computation model, which makes it a good choice for real-time and dynamic environments. The principle of the APF approach is based on attracting the MR to its target position and repelling it away from obstacles. The APF approach has some inherent limitations, such as local minima and path oscillations in narrow passages. This paper reviews the traditional APF approach and several modified versions that aimed to overcome these limitations. A brief comparison is made to give an overview of the performance of each modified version. In addition, the main advantages and disadvantages are discussed to highlight the parameters affecting the performance of the modified methods.