This paper presents the first application of a higher-order Smoothed Particle Hydrodynamics (SPH) method to the thermal modeling of metal cutting problems. With this application, the heat transfer equation in the thermo-mechanical simulation of metal cutting is solved more accurately by addressing the consistency issue of standard SPH formulations. Furthermore, through a robust and effective surface-detection algorithm, this work enables the SPH cutting models to include heat loss thermal boundary conditions for the first time. Process forces, tool temperatures, and chip geometry are numerically investigated in machining a Ti6Al4V workpiece at two different cutting speeds. Several validation tests and sensitivity analyses are performed in high resolution, thanks to the runtime acceleration of SPH by parallel computing on Graphics Processing Units (GPUs). The results show that SPH simulations with the proposed thermal modeling approach achieve more realistic serrated chips in titanium cutting problems.