In the present work, the welding process was using the tubular wire process with pulsed current and gas protection in flat position, SAE 1020 steel base metal and addition metal the martensitic stainless steel 410 NiMo (CA6NM), which has intensive turbine recovery. The purpose was to optimize the welding parameters (average current, pulse rate, welding speed and part contact nozzle distance) in the coating, assisted by the use of first order regression equations in the response variables (process yeld, width, reinforcement and reinforcement area) to predict the results. For the design of the experimental matrix with four variables and three levels, a TAGUCHI method was applied, and the reconstruction of 3D images with a laser device and a web camera, that was adjusted to measure the geometry of the weld bead.The analysis of variance was applied in the results obtained, to statistically prove which variable exerted influence on the response variables, observing that the best performance proceeded for the average current of 230 amps, since statistically the average current was one of the variables of influence that most affected the responses. Therefore, the results of the confirmatory tests with the optimized parameters, showed that the developed models are able to predict the response variables with precision, demonstrating the effectiveness of the TAGUCHI optimization method. The current and voltage signals were obtained simultaneously, and using the Fourier analysis the changes in signal behavior, with frequency and time information that generated the electric arc stability were identified.