Friction stir welding (FSW) is a non-traditional welding procedure for joining identical and divergent substances that are comparatively hard to weld using fusion techniques. The FSW is more competent and eco-responsive when correlated with the traditional processes. The tensile characteristics of thermally susceptible alloys are minimized as a result of heat iterations. The roughening of the consolidated deposits in friction stir welding induces joint moderation. For non-heat-treatable alloys, intergranular and stress corrosion cracking are developed in grain boundaries during the FSW. Because of this, the overall performance and mechanical properties were reduced. Underwater friction stir welding (UWFSW) is preferred to prevail over the problems. UWFSW is appropriate for both heat-treatable and non-heat-treatable alloys that are receptive to thermal activity throughout the procedure. This research examines the effects of response surface methodology (RSM)-based grey relational methods on the optimization of UWFSW procedural characteristics such as the straight hexagonal tool profile (SH), tool rotational speed (TRS) of 1200 rpm, tool transverse velocity (TTV) of 20 mm/min, and water head (WH) of 10 mm in military-grade AA5083 alloy. The mechanical properties of the AA5083 UWFSW joint, such as its average ultimate tensile strength (UTS) and hardness, have been greatly improved. Grey Relation Analysis was used to determine welding specifics such as UTS and hardness. The ANOVA was used to assess the formulation's significance. The micro-structural behaviour and grain structure of different weld zones were investigated with the help of the scanning electron microscope (SEM). The image processing technique can be used to quantitatively assess the UWFSW process using the non-destructive testing (NDT) ultrasonic B-scan image and radiographic image. Confirmation tests at the optimal parameter level improved the performance of each response. © IMechE 2022.