In times of rising prices for raw materials and the worldwide need of reducing energy consumption, the construction of lightweight components is more relevant than ever. Therefore the use of carbon fiber reinforced polymers plays an extraordinary role and a broad understanding of their properties is of great importance. This thesis deals with fatigue delamination testing of unidirectional carbon fiber reinforced laminates with both, epoxy and polyetheretherketone (PEEK) as matrix materials, under mode I and mode II loading conditions. The goal for the mode I fatigue tests was to develop a test method to measure the crack propagation rate in a Double Cantilever Beam specimen at a constant strain energy release rate. Therefor in a first approach the crack tip opening diplacement was increased step-wise using predefined levels of machine displacement according to certain increments of crack length. This method was evolutionary improved until the energy release rate was held at a constant value by real-time processing of the measured load and displacement data based on a preliminary compliance calibration. This approach yielded the best results but has the problem that by maintaining the maximum and minimum energy release rate constant, the R-ratio of the displacement varies. Based on this problem, further investigations to improve this method might be performed. The basic idea for the mode II fatigue tests was to compare the tests based on the End Notched Flexural and Calibrated End Loaded Split configuration. Especially the applicability of the Calibrated End Loaded Split test under fatigue loading was the target of investigation. Therefore an End Loaded Split fixture was constructed and first preliminary tests were performed. In a second stage material for an international round robin for the Technical Committee 4 of the European Structural Integrity Society was tested and the results are presented in this thesis. The tests yielded different results in the plot of energy release rate vs. crack propagation. No effects due to the mounting of the specimens could be proved by measurements of the local strains and heating of the specimen, but the reason for the different results could not be found yet. It will be interesting to see if the results of the other laboratories participating in the round robin show the same tendency. If they do so, the applicability of both test setups for measurements under fatigue loading conditions have to be questioned and further research might be performed.