© 2014 Elsevier Ltd. All rights reserved. Glass is an omnipresent material which is widely used as façade in buildings. Damage of glass windows and the associated glass fragments induced by impact and blast loads impose great threats to people in the vicinity. Much effort has been directed towards understanding glass material properties, and modeling of glass window responses to impact and blast loads. For reliable predictions of glass structure performances under dynamic loadings, an accurate dynamic constitutive model of annealed float glass, which is commonly used for glass windows, is therefore needed. In current practice, the Johnson-Holmquist Ceramic (JH2) model is most commonly used in simulating glass plate responses to impact and blast loads. In this study, the accuracy of the JH2 model in modeling annealed float glass material, especially at high strain rate is examined in detail. Static compressive tests and dynamic compressive tests using Split Hopkinson Pressure Bar (SHPB) are carried out on soda-lime glass specimens sampled from commercially used annealed float glass panes. These testing results are used together with the authors' previous testing data and data reported by other researchers in the literature to determine the constitutive constants for the JH2 model, including Equation of State (EOS), strength criterion and strain-rate effect. The JH2 model with new material constants is then programmed in commercial code LS-DYNA. To verify the model, it is used to simulate a SHPB compressive test on a 15 mm by 15 mm (diameter by length) glass specimen, a field blasting test on a laminated glass window of 1.5 m by 1.2 m in dimension, and a full-scale laboratory windborne debris impact test on a laminated glass window. The simulation results demonstrate that the JH2 model with the new material constants for annealed glass gives good predictions of glass material and glass window responses to impact and blast loads.