In-space service, assembly, and manufacturing (ISAM) promises to accelerate the development cycle by deploying the tools to update existing systems or to create new systems in space. The space environment offers opportunities and challenges to perform ISAM activities. A fully realized vision for ISAM would also enable larger scale and new concepts, exploiting the microgravity environment to achieve large work volumes and novel structural concepts. On the other hand, it expected ISAM to be autonomous and, in some cases, with machine learning capabilities, but in space, computational and electric power are more limited than on Earth. For instance, the avionics and onboard computers would suffer overheating and throttling due to the vacuum environment. Also, atmospheric drag and gravitational attraction can accumulate to the level of challenging station keeping and proximity operations. Therefore, optimal and intelligent approaches are required to enable autonomous and intelligent ISAM activities. Digital twins (DT) offer a solution to this challenge by creating virtual representations of the physical systems and maintaining information about the configuration and state of these systems, facilitating and accelerating the analysis of different ISAM scenarios under various space environment conditions. However, due to the space environment characteristics, for the DT to run in real-time, proper selection of the data and the correct level of fidelity is essential to ensure feasibility. The proposed project will: develop a digital engineering framework to create a digital thread of a robotic- based ISAM system and process and the corresponding DT; introduce methodologies and model order reduction techniques to determine an optimal level of fidelity of DTs dynamically; optimally select a subset of available data to be ingested by the DT estimation algorithm; assess the impact of the space environment on ISAM activities.