The early integration of circular strategies into product development is crucial for contributing to a sustainable circular economy. Since complex product developments rely on modeling, it is reasonable to integrate circular strategies into these models. This integration, however, poses economic challenges due to the absence of modeling standards for non-linear lifecycles and the increased modeling effort required. This paper introduces a structured methodology for integrating circular strategies into Model-based Systems Engineering (MBSE) while reducing manual modeling efforts. The proposed methodology utilizes initial MBSE models tailored to specific circular strategy subcontexts. Semantic interoperability between the Systems Modeling Language (SysML) and Kernel Modeling Language (KerML) is used to provide a link to Large Language Models. Retrieval-Augmented Generation is then used to formalize recommendations for the subcontext model. These recommendations are provided to the developers via KerML within the SysML environment, assisting them to reduce the required manual effort for subcontext modeling. The evaluation of a subcontext expansion demonstrates a reduction in modeling effort and an increased information density while maintaining model consistency and quality. Furthermore, the iterative recommendation mechanism supports continuous model evolution, making this approach particularly valuable for implementing circular strategies in adaptive, iterative product development.