As the demand for solving complex problems, such as multiphysics, multiscale models, and real-time simulations, grows, efficient solvers are essential. Domain decomposition (DD) techniques and reduced-order models (ROMs) have emerged as key tools for addressing these challenges. DD solvers decouple subcomponents, enabling the integration of state-of-the-art numerical codes, while ROMs facilitate fast computations, especially for real-time or multi-query applications. These methods are particularly valuable for multi-physics problems where efficient subcomponent solvers are available or where direct access to numerical algorithms is restricted. This talk introduces a comprehensive optimisation-based DD framework for coupling diverse discretisation methods in Computational Fluid Dynamics (CFD), maintaining physical consistency across domains. I will present optimal control DD formulations and strategies for selecting effective optimisation algorithms to decouple subdomain computations. Additionally, I will showcase the integration of intrusive and non-intrusive model-order reduction approaches, leveraging the efficiency of data-driven methods while maintaining the accuracy of projection-based ROMs. Finally, I will discuss heterogeneous couplings, focusing on FEM-ROM DD interactions, while also highlighting recent extensions to multi-physics problems and the incorporation of time-adaptive strategies for enhanced efficiency.