The characterisation of materials in thermomechanically (thermomechanical fatigue, TMF) highly stressed engine parts (e.g. cylinder head) is usually tested on uniaxial and cylindric-shaped specimens. With this testing method actually occurring spatial temperature gradients are not recorded. This results in a multiaxial stress state which is exacerbated by the complex part geometry. Multiaxial material testing is often carried out by testing the assembly of the internal combustion engine. To efficiently optimize materials it is reasonable to close the gap between uniaxial TMF-testing and assembly-testing. In conventional TMF-testing, cylindric specimens are inductive heated. The benefit of this technology ist the contactless and controllable heat input. These advantages motivate the use of induction coils for heating complex shaped parts. To be able to optimize the coil geometry (and thereby the temperature field) in the future, a simulation process has been developed in this diploma thesis. The aim is to introduce an accurate spatial and temporal temperature distribution into the component. For this purpose the heating of uniaxial TMF-specimen, in use of the method of finite elements (FEM), has been modelled and calibrated to measured temperatures. In the next step the simulation has been transfered and varified to a more complex geometry (plate with two drill holes). The accordance between simulation an testing has been confirmed in both cases. In the last step the temperature distribution of a cylinder head has been calculated numerically. The temperature field has been transferred to a further software for a concluding mechanical simulation. Thus, the foundation for the development of a forward-looking test methodology was laid.