Many destructive test methods are used to prove the resistance against loads occurring in a bonded joint during its life cycle. In addition to the determined mechanical properties, the evaluation of the failure behavior is an important quality criterion. Maximum adhesive strength is achieved with cohesive failure and this joint strength can be predicted. Based on this fact it is more important to determine the failure percentages objectively and reproducibly. In addition, adhesive-specific, characteristic structures are formed in cohesive fracture surfaces which indicate the failure-relevant, locally acting loads. In industrial practice, there is no evaluation method to objectively evaluate and analyze the failure ratio of an adhesively bonded joint. The innovation of the developed, computer-aided analysis of the fracture surfaces involves the objective determination of the failure ratio according to international standards and the further analysis of the failure-relevant fracture processes. This enables the adhesive user to evaluate a large number of test specimens in a time-efficient, objective and automated manner. The automated evaluation method calculates the failure mode proportions based on digital images of the fracture surface. The segmentation and classification use a histogram-based color differentiation. The region growing algorithms were optimized and tested with a database of different adhesives and materials. Furthermore, it is possible to get knowledge about load conditions of the destroyed joint by analyzing the fracture surface. To quantify the relationship between load conditions and the resulting fracture structure a topographical 3d method and a 2d imaging method is investigated. The differentiation between the load angle and testing temperatures shows significant, adhesive-specific characteristics.