Cyber-physical systems (CPS) are heterogeneous, because they tightly couple computation, communication, and control along with physical dynamics, which are traditionally considered separately. Without a comprehensive modeling formalism, model-based development of CPS involves using a multitude of models in a variety of formalisms that capture various aspects of the system design, such as software design, networking design, physical models, and protocol design. Without a rigorous unifying framework, system integration and integration of the analysis results for various models remains ad hoc. In this paper, we propose a multi-view architecture framework that treats models as views of the underlying system structure and uses structural and semantic mappings to ensure consistency and enable system-level verification in a hierarchical and compositional manner. Throughout the paper, the theoretical concepts are illustrated using two examples: a quadrotor and an automotive intersection collision avoidance system.