Welding is the process of uniting two or more metals that employed in practically every engineering application. In fasil engineering for structural steelwork and truck trailers components are welded manually which in slower and with in a human error, resulting welding defact, non uniform, and less quality. All these processes are performed through Metal inert gas welding. While, there are different welding defects like porosity, spatter, and undercut were encountered. Through these defects, tensile strength, and hardness of the weld were dramatically impacted. Since, this company welds without considering the optimum values of the welding parameters. As a result, this study has provided that process parameters optimization of MIG welding on mild steel AISI 1020 to optimize input parameters of welding current, welding voltage, welding speed, and shielding gas with their tensile strength, and hardness. Taguchi 𝐿9 an orthogonal array was employed in the design of the experiment. The objective function of a hybrid artificial neural network with in Genetic algorithm (ANN-GA) was implemented. This model was used to predicted and optimize UTS and hardness. The results of the ANN-GA model could forecast the output responses with a mean square error of 0.312597e-1 with in a 4-8-2 network trained of back propagation, and Levenberg Marqudrt learning aligorithm.The optimal set of design variables were obtained like welding current 112.5A, welding voltage 28.6V, gas flow rate 20L/min, and welding speed 175.8mm/min were found. Along with their response values of ultimate tensile strength 434.9 MPa, and hardness 69.78 HRB. Finally, from a confirmation tests, the average result of 429.23 MPa of UTS, and 68.8 HRB of hardens were obtained. The percentage errors of ANN-GA predicted optimal responses results and the confirmatory results were found 1.25% and 1.79% for UTS and hardens. Considering that, the optimization process was a minimal % of error and an acceptable level of correctness. Keywords: Metal Inert Gas Welding, AISI 1020 mild steel, Artificial neural network, Genetic algorithm, Mechanical properties.