Abstract: Objectives: While exploring the vehicle-road-cloud integration and automotive intelligence, it turns out that spatial-temporal data of Intelligent Connected Vehicles (intelligent connected vehicle, ICV) has emerged as a crucial data element in driving productivity and technological transformation. Spatialtemporal data is highly sensitive, multi-source, heterogeneous, and dynamically circulated. It is often necessary to set up governance mechanism for ICV spatial-temporal data. Currently there is still no modelling framework for spatial-temporal data governance and the mechanism is missing. Methods: A data governance model concerning ICV spatial-temporal data is proposed by extending the governance model from the general information domain. It highlights the spatial-temporal data governance entities, governance bodies, and governance environments as core entities in the modelling. It also defines six basic relationships among core entities. Leveraging the model and provenance technologies, the paper presents a reference framework for governance implementation across various stages of ICV spatial-temporal data lifecycle. Furthermore, an empirical study is performed on representative cases of spatial-temporal data governance from the ICV industry. Results: By applying the governance model in the ICV domain, the results provide several recommendations for the governance of ICV spatial-temporal data: (1) Governance cooperation between automotive enterprises and authorized map service providers. This option leverages the existing qualifications and rich experience of map service providers to offer automotive enterprises a quick way to meet governance requirements. (2) Internal governance by automotive enterprises. The option provides flexibility for automotive enterprises to build spatial-temporal data governance workflow, yet it requires the automotive enterprises to get authorization for map production. (3) National third-party governance. The option allows the government to authorize third-parties to enforce governance, and create a national spatial data infrastructure to support the governance. These recommendations provide a variety of options for different enterprises to choose suitable approaches based on their specific circumstances, paving the way for more inclusive and effective governance approaches. Conclusions: The proposed model and reference framework comprehensively consider various stakeholders and practical cases in the governance system, and provides a theoretical foundation for developing a governance system for ICV spatial-temporal data. It also shows that the development of a national crowd-sourced mapping infrastructure covering maps and scene libraries is a promising direction for both data governance and automated driving.