The MRO industry faces substantial challenges with regard to the capacity planning of disassembly and reassembly work. This is due to the unknown workloads when regenerating complex investment goods and is caused, in particular, by the uncertain degree of disassembly and the complex challenges of reassembly. Forecasting techniques based on Bayesian networks are developed along with mathematical models which optimize capacity utilization, job order and the resulting costs. The approaches are tested and validated in conjunction with an MRO company with global operations. The results show possibilities for enhancing the planning processes and are found to be transferable on an international scale regardless of sociocultural and process differences.