This work addresses the problem of controlling the width of a weld bead in the CMT process in the context of additive manufacturing. The control design is carried out considering a linear model with delay and unknown gain. An MRAC adaptive control approach was used in conjunction with a Smith Predictor to control the weld bead width, which is estimated using a computer vision-based strategy. The proposed control is numerically simulated and shows satisfactory results with tracking and parametric errors converging. In addition, an integral control has been implemented in the welding cell in order to obtain preliminary closed-loop control results and verify 
the feasibility of bead geometry control.