The association of system-level heterogeneous models is a core component of the integration between PLM and MBSE, and it is a key step in facilitating the application of MBSE in the digital collaborative R&D of complex equipment. To address bottlenecks such as difficulties in adapting heterogeneous toolchains, insufficient formal expression of business rules, and missing interoperability interfaces, this study proposes a model association framework based on SysML v2 semantic constraints. This framework leverages the precise formal modeling capabilities of SysML v2 to achieve mechanisms for defining association rules, feature transformation, and constraint validation across domain models. It constructs a unified interoperability environment that supports browsing, associating, transforming, tracing, and building rule libraries for heterogeneous models throughout their entire lifecycle. The framework standardizes interfaces to achieve deep integration between PLM and MBSE, supporting the formal expression of various association rules and consistent parameter transmission in collaborative modes. Additionally, it lays the foundation for model interoperability to drive AI-driven process optimization and knowledge reuse in R&D.