The flow stress in the high-speed machining of titanium alloys depends strongly on the microstructural state of the material which is defined by the composition of the material, its starting microstructure, and the thermomechanical loads imposed during the machining process. In the past, researchers have determined the flow stress empirically as a function of mechanical state parameters, such as strain, strain rate, and temperature while ignoring the changes in the microstructural state such as phase transformations. This paper presents a microstructure-sensitive flow stress model based on the self-consistent method (SCM) that includes the effects of chemical composition, α phase and β phase, as well mechanical state imposed. This flow stress is developed to model the flow behavior of titanium alloys in machining at speed of higher than 5 m/s, characterized by extremely high strains (2–10 or higher), high strain rates (104–106 s−1 or higher), and high temperatures (600–1300 °C). The flow stress sensitivity to mechanical and material parameters is analyzed. A new SCM-based Johnson–Cook (JC) flow stress model is proposed whose constants and ranges are determined using experimental data from literature and the physical basis for SCM approach. This new flow stress is successfully implemented in the finite-element (FE) framework to simulate machining. The predicted results confirm that the new model is much more effective and reliable than the original JC model in predicting chip segmentation in the high-speed machining of titanium Ti–6Al–4V alloy.