The study aims to analyze and compare A* and AO* heuristic-based path finding algorithms for a multi-weighted goals traversing in a 2-D grid. The focus is to determine which algorithm can be optimized and efficiently implemented. Both algorithms are trained using PER reinforcement learning methods. The grid environment is twenty percent covered by obstacles. It constitutes a starting point and five Goals, which are traversed based on their pre-assigned priority. When the prior Goal has been successfully traversed, the shortest path to the subsequent Goal is determined from the current point. While the heuristic techniques are compared, the environment setup and the RL method remains the same. The algorithms are tested in a grid of dimensions 20 × 20 and each one is trained for a thousand episodes. The Analysis and the comparison is based on the calculated loss, the success rate and the path length. The average loss of A* and AO* calculated after training them for 1000 episodes was 0.0003 and 0.3960 respectively. While the average path length is 51.77 units for A* and 47.97 units for AO*. It was concluded that the A* heuristic algorithm performed much better than the AO* in the present environment.