In order to solve the dual-objective path optimization problem of arc welding robot, a complete and novel methodology is proposed, including modeling, collision-free path search and global path optimization. First of all, the grid method is utilized to accurately model the welding workpieces and surrounding environment. Afterwards, an adaptive extension bidirectional RRT* algorithm (AB-RRT*) is used to search for collision-free paths between any two weld seams. Finally, aiming to minimize path length and energy consumption, the global path is optimized through the improved discrete NSGA-III algorithm (IDNSGA-III). Simulation results show that the AB-RRT* algorithm can search for optimal or sub-optimal solutions with higher efficiency. The IDNSGA-III algorithm can properly maintain the trade-off between convergence and distribution. Furthermore, it can obtain satisfactory solutions in a shorter time and provide effective guidance and help for engineers to plan the actual path.