Abstract: This paper takes the shape of single weld as the starting point to study the influence of welding process parameters on the weld bead morphology and establish a regression equation to predict the weld bead size. The research shows that the welding current has the most significant effect on the weld bead size. When the welding current increases by 50 A, the weld width increases by 14% on average. There is a certain coupling relationship between different process parameters. At the same time, it reveals the law that the increase of welding current and welding speed causes the change of arc dynamic pressure, which leads to the transformation of weld shape from finger to bowl. The reason for the change from finger-like morphology to bowl-like morphology. Compared with various regression models, the response surface method based on the elimination of low correlation factors proposed by considering the coupling effect of process parameters has a good overall regression effect, with R² above 0.97 and the average error within 0.5%. The regression model established in this paper will provide accurate weld bead size information for the planning of arc surfacing layers and passes, and ensure that the robot welding pose meets the process requirements.