In this study, we conducted a low-carbon-oriented performance comparison of three mainstream path planning algorithms, A-star (A*), Rapidly exploring random tree (RRT), and Rapidly exploring random tree-star (RRT*), under a unified platform. Through image-based map simulation and obstacle inflation processing, a consistent experimental environment is established with fixed start and goal points set to ensure comparability. The experimental evaluation metrics include total path length, algorithm computation time, node expansion behavior, and path smoothness. The results show that A* performs best in structured environments, producing the shortest path. Although RRT has fast exploration capabilities, it tends to generate irregular trajectories. RRT* improves path quality through a node rewiring mechanism, making it suitable for carbon-sensitive scenarios, but with higher computational cost. Overall, the results of this study fill the gap in previous carbon-sensitive navigation research by providing a cross-algorithm comparison. It offers an empirical foundation and visual reference for selecting appropriate strategies in future low-carbon autonomous mobility systems.