The understanding of mechanical behavior in machining is critical to analyze and design a process. It is well known that work materials experience large strains, high strain rates, high temperatures, and complex loading histories. Adiabatic or quasi-adiabatic condition is an important feature of material deformations in machining ferrous alloys. The problem of how to accurately model the mechanical behavior including the adiabatic effect is essential to understand a machining process. Several constitutive equations such as the simple power law model, Johnson-Cook (JC) model, and other models have often been used to approximate flow stress in machining analysis and simulations. The JC and other empirical or semi-empirical models lack mechanisms in incorporating complex loading effects. The internal state variable plasticity Baumann-ChiesaJohnson model (BCJ model) has been shown to incorporate loading histories as well as state variables. In this study, we have determined the material constants of AISI 52100 steel (62 HRc) for both the JC and BCJ models using the same baseline stress-strain data. The material constants were obtained by fitting the JC and BCJ models to these test data at different strains, strain rates, and temperatures using nonlinear least square methods. Both models are capable of modeling strain hardening and thermal softening phenomena. However, the BCJ model can also accommodate the adiabatic effect, while the JC model is basically isothermal. Orthogonal cutting tests and FEA simulations, based on the design-of-experiment method, were performed using the cutting tool with a 20° chamfer angle. The predicted saw-tooth chip morphology and dimensions using the BCJ model are consistent with the measured chips in the cutting tests, while the JC model yielded discontinuous chips. In addition, the BCJ model gave larger subsurface von Mises stress, plastic strain, and temperature compared with those by the JC model.