In order to conduct uniaxial tensile tests for hot stamping applications, tests are normally performed by using a Gleeble thermo-mechanical materials simulator so that rapid heating and cooling processes can be obtained. However, temperature gradients in a specimen tested on Gleeble are inevitable due to resistance heating principles and heat loss to grips and water-cooled jaws. In this research, a pair of purpose-built grips made of stainless steel with low thermal conductivity and significantly reduced contacting area for clamping, as well as a flat dog-bone specimen with maximised parallel length (80 mm) were designed, for the purpose of improving the temperature uniformity within the concerned gauge section area of the specimen. Uniaxial tensile tests on AA6082 were performed, after controlled heating and cooling processes, at constant deformation temperatures in the range of 400 ℃–500 ℃ and at constant strain rate in the range of 0.1–4/s, to simulate its hot stamping conditions. The digital image correlation system was adopted to enable strain distributions in specimens to be measured. The temperature distributions in specimens were investigated and an effective gauge length of 14 mm was specified accordingly to ensure temperature gradients less than 10 ℃ within it at all tested temperatures. True stress–true strain curves of AA6082 were obtained based on results of strain measurements along the defined effective gauge length and used to calibrate a set of advanced material model. Error analysis was carried out by using thermo-electrical and thermo-mechanical FE models on ABAQUS, in which the calibrated material constitutive equations were implemented via subroutines. The error of stress–strain curves of AA6082 measured based on the specified gauge length was investigated and quantified by analysing the distribution of axial strain and axial stress.