A material softening model based on thermal activation energy has been successfully established through tempering experiments in the first part of this study. To apply the model to predicting material softening in hard turned surfaces, the thermal history of work material is needed. In this part, a three-dimensional finite element (FE) model of machining hardened 52100 steel is constructed, and coupled thermal-stress analysis is performed to obtain the material thermal history. Then the material softening model uses the computed thermal history as input to predict the material hardness profiles along the depth into the machined surfaces. Overall, the prediction precisely catches the trend of hardness change along depth and agrees reasonably well with the hardness measurement. What’s more, the sensitivity of material softening to cutting parameters is investigated both quantitatively and qualitatively. Within the investigation range, it is observed that the increase of tool flank wear and feed rate produces severe material softening and a deeper softened layer, while the increase of cutting speed causes significant softening to the surface material but hardly changes the softened depth.