Cyber-physical systems are highly collaborative by nature. At runtime these systems collaborate with each other to achieve goals that a single system could not achieve on its own. For example, autonomous vehicles can dynamically form convoys at runtime to facilitate higher traffic throughput and a reduction in CO2 emissions. While the importance of context documentation and analysis in system development is well known, current model-based engineering approaches struggle with the size and complexity of cyber-physical systems’ contexts. This is due to high variety and dynamicity of the contexts to be considered. For example, a convoy to be formed at runtime may consist of different numbers of participating vehicles. Additionally, it may face different neighboring, not partaking context systems (e.g., non-equipped vehicles, equipped but not participating vehicles) and situations (e.g., speed limits, road construction sites, emergency vehicles). This paper proposes a context ontology to cope with highly dynamic contexts of cyber-physical systems by explicitly differentiating between not only the system and its context but also between the cyber-physical system network the system participates in, as well as the system network’s context.