Tension and compression testing has been performed over a wide range of temperatures (23–250°C), strain rates (0.001–1s−1) and material orientations to characterize the mechanical response of AZ31B-O. Instantaneous r-values were measured and their evolution with plastic strain. A strong evolving asymmetry/anisotropy is observed comparing the tension and compression flow stresses and r-values at room temperature, while weaker anisotropy/asymmetry is seen at higher temperatures. Higher strain rate sensitivity is measured at elevated temperatures. An evolving anisotropic/asymmetric continuum level material model based on a CPB-type yield function (Cazacu et al., 2006), as modified by Ghaffari Tari et al. (2014), was used to fit the material behavior at different temperatures. Two approaches are considered to introduce thermal softening and strain rate sensitivity. For isothermal conditions, a rate-sensitive Cowper–Symonds strain hardening model, fit at individual temperatures, captured the material behavior well. For non-isothermal conditions, a strain rate and temperature dependent hardening response based on a modified Nadai model was used. Both approaches have been coded into a user material subroutine within the commercial finite element package, LS-DYNA and used to simulate limiting dome height experiments performed at 250°C and non-isothermal deep drawing experiments in which the temperature distribution within the blank ranged between 170°C and 250°C.