Abstarct: ABSTRACT One of the new methods for solid-state joining of aluminum alloys is friction stir welding (FSW). In this research, the process of designing, manufacturing and testing the rotary friction stir welding (RFSW) equipment is presented, which is used to join Al6061 pipes in the butt position and with the possibility of placing alumina reinforcing microparticles. For this purpose, statistical analysis and optimization of input variables such as tool rotational speed, welding speed and the size of reinforcing microparticles have been done to evaluate the mechanical properties of joints using the principles of DOE. In this regard, the response surface methodology (RSM) and Box-Behnken design (BBD) were chosen as the experimental design method, and the optimal conditions of input variables were extracted using the desirability function. Maximum tensile force, maximum engineering stress, maximum engineering strain, MHD parameter and heterogeneity index were selected as response parameters. In order to quantitatively and qualitatively evaluation of the mechanical properties of joints, tensile, microhardness and mextallographic tests were performed on the produced samples. The results of analysis of variance (ANOVA) showed that the tool rotational speed has an effect on all the response parameters in the form of first and second order terms. Meanwhile, welding speed has a significant effect on MHD parameter and heterogeneity index in the form of first and second order terms. The values of the desirability function to maximize tensile force, engineering stress and engineering strain are respectively equal to 90.2, 92.4 and 100% and the values of the desirability function to minimize the MHD parameter and the heterogeneity index were extracted respectively equal to 86.1 and 76.2%. In order to verify the optimal combination of input variables, experimental tests were performed using optimal conditions. The difference percentage of the results obtained from the optimization process and verification tests for the parameters of maximum tensile force, maximum engineering stress, maximum engineering strain, MHD parameter and heterogeneity index are equal to 2.91, 3.12, 3.12, 9.23 and 7.81% which indicates the ability and good performance of regression models to predict the response parameters.