Abstract Process induced heat causes workpiece shape deviations during machining and may result in rejected parts. As unpredictable processing and boundary conditions cannot be simulated in the process planning phase, a process parallel simulation of the workpiece temperature is required to observe and compensate thermal effects. To parametrize the process parallel simulation model, a parameter identification method for milling operations is developed. With model order reduction (MOR) the computational time is reduced to enable the process parallel simulation. An observer based on a few temperature measurements during machining is designed to reconstruct the non-measurable temperature distribution. A validation proves the methods’ potential.