Split Hopkinson pressure bar (SHPB) tests are popular in calibrating material models at high strain rates under the assumption of constant strain rates. Nonetheless, numerous strain rates obtained from the SHPB test cannot be approximated as constant. Here, the constant assumption is discarded and a new calibration approach with history data is proposed and illustrated with the Kelvin-Voigt model. Intensive SHPB tests on aluminum alloy LY12 and red copper were carried out to obtain time history data and to validate the new calibration method. Those time history test data published in the literature were also utilized to validate the method for the Kelvin-Voigt model. The results show that the SHPB test data from a variety of materials can ideally fit the model. • It is emphasized that the behaviors of materials at high strain rates have rheological characteristics. • Discarding the constant assumption and using historical data to calibrate models for materials at high strain rates. • The Kelvin-Voigt model is suitable for a variety of materials at high strain rates, which is validated by SHPB tests of YL12 and red copper as well as SHPB test data of four materials from the literature. • If the shape of the reflected wave is far from a rectangle, the constant strain-rate assumption arises significant errors.