In order to achieve single pass welds without edge preparation, instead of multipass procedures, one of the most notable techniques is to use activating flux with GTA welding (GTAWflux) process. In autogenous GTA welding mixed fluxes (oxide SiO2 and Cr2O3), were applied on the weld area of type 310 stainless steel through a layer of the flux to produce a bead-on-plate weld. Many parameters affect the GTAW-flux process quality. The Taguchi method can improve on the disadvantages of full factorial design. However, in practice, it has certain limitations. In this work, a Taguchi-Neural approach (Combine Taguchi method and a neural network) was used to select the GTAW-flux process parameters. The experimental procedure of GTAW-flux process, via the proposed approach produced full penetration of the weld root in 8 mm thick type 310 stainless steel with single pass welds.