In this paper, rolled sheets of aluminum alloy 5052-H18 were joined by friction stir welding. Several samples were welded by varying the principal parameters including tool rotational speed, linear speed, tool shoulder diameter, and pin diameter. Response surface methodology (RSM) was used as the statistical design of experiment technique for parameter optimization. A predictive model was developed to determine optimum parameters that maximize tensile strength. Results showed that the tool rotational speed was the most important parameters affecting the tensile strength of aluminum alloy 5052 joints. Maximum tensile strength was achieved using the rotational speed of 1233 rpm, the linear speed of 107 mm/min, the shoulder diameter of 12.8 mm, and the pin diameter of 1.8 mm. The difference in tensile strength between the predicted value and the measured value was 2.1%. The effect of tool rotational speed and linear speed on the size of the heat-affected zone was investigated. Consequently, a three-dimensional model was developed to predict the size of heat-affected zone and the temperature history of the material during the FSW process and the results were compared with experimental data.