Abstract: When welding in space with robot, the efficiency, smoothness, and safety of path planning are key factors affecting welding quality and robot performance. To this end, a welding path planning method based on an improved ant colony algorithm is proposed. Firstly, a spatial geometric model of the welding workpiece is established, and key points of the weld path are extracted using an adaptive discretization strategy. Secondly, heuristic information and pheromone update mechanisms are introduced, combined with a dynamic visibility factor to enhance the globality and convergence of path search. Finally, through smoothness constraints and multi-objective optimization functions, comprehensive optimization of welding path length, safety, and smoothness is achieved. Simulation results show that the proposed method outperforms traditional planning algorithms in terms of path length and smoothness, with an average reduction of about 18.7% in total path length, and good trajectory continuity, meeting the practical requirements of industrial welding. Welding experiments further verify the feasibility and effectiveness of this method. The welding path planning method based on the improved ant colony algorithm has significant advantages in improving welding efficiency and quality, and is suitable for automated welding tasks with complex weld structures.