Abstract: Laser welding process monitoring is one of the key aspects in the intelligent laser welding process. This study investigates the feasibility of using in-situ monitoring of the molten pool's mirror-like microregion to predict weld seam surface formation quality. By controlling the shielding gas flow rate, we alter the molten pool surface's mirror-like microregion characteristics, and analyze the correlation with post-weld surface morphology. Results show that the surface roughness of the weld decreases and then increases as shielding gas flow rate increases, with the minimum occurring at a flow rate of 5 L/min. The total area of the molten pool’s mirror-like microregion first increases and then decreases with changes in the shielding gas flow rate: at low welding speeds, the maximum value of the mirror-like microregion area is increased by about 90% compared to no shielding gas, while at high welding speeds, it increases by about five times. The ratio of the total area of the mirror-like microregion to the total molten pool surface area increases and then decreases with the increase in shielding gas flow rate, which is negatively correlated with the surface roughness of the weld. Further analysis suggests that monitoring the behavior of the mirror-like microregion area in the molten pool surface in real-time could provide a basis for online prediction of weld seam surface formation quality.