The main aim of this thesis is the systematic characterisation of the fatigue behaviour of short glass fibre reinforced plastics with Wöhler fatigue tests and the generation of suitable hypotheses for fatigue life calculation of dynamically loaded components made of such materials. The basis for dimensioning of dynamically loaded components concerning the fatigue life is the knowledge of the cyclic material behaviour considering influences on fatigue life which are relevant for the respective application. To generate a comprehensive database, three short glass fibre reinforced polymers are characterised in detail. The focus is on the specification of influencing parameters on Wöhler curves such as type of matrix, fibre orientation, stress ratio, temperature, environmental media, weld lines and notches. Hypotheses for fatigue life prediction of reinforced polymers are derived from a lot of tests on specimens. These allow for a transfer of the material data generated on simple specimen geometries to complex component structures. For fatigue life calculation of short glass fibre reinforced polymers these hypotheses are implemented into a standard fatigue software tool FEMFAT and validated with component tests. For fatigue life prediction a closed simulation process was established, which takes into account the fibre orientation and distribution as a result from an injection moulding simulation. Suitable software and model interfaces have been developed for this purpose. Moreover, for structural stress analysis with the finite element method the local anisotropic material behaviour can be considered. Therefore, the injection moulding simulation supports the process chain as an additional tool. For the fatigue life calculation of dynamically loaded complex component structures made of short glass fibre reinforced polymers a closed simulation chain consisting of injection moulding simulation, stress analysis with the finite element method taking into account the anisotropic material behaviour and finally the fatigue life prediction can be realised. Hence, in this dissertation for the first time a procedure is presented that considers polymer specific influences in the product development process in order to improve time and cost efficiency.