Vehicle system testing serves as a critical phase in obtaining road certification for prototype vehicles. While direct road testing with physical vehicles yields the most authentic data, this approach entails significant costs, challenges in reproducing extreme scenarios, and inherent safety risks. In contrast, virtual vehicle-based testing technologies represent advanced simulation methodologies for enhancing development efficiency and quality, effectively mitigating risks associated with complex real-world operating conditions and hazardous physical testing. However, virtual vehicle models often rely on idealized parameters, limiting their ability to reflect real-world dynamics and resulting in lower credibility of test outcomes. Furthermore, as evidenced in current mainstream virtual testing software, environmental simulations predominantly remain confined to the visual domain, with limited direct interaction between dynamic environmental changes and virtual vehicle responses. To address these limitations, this study proposes a novel testing framework leveraging vehicle-cloud integration technology, which combines the authenticity of physical testing with the flexibility of virtual simulation. The proposed system is validated through an AEB (Automatic Emergency Braking) function activation test. Experimental results demonstrate real-time data interoperability between physical and virtual vehicles, achieving a 89% accuracy rate in synchronizing virtual scenario velocities with real-world speeds. This approach enables safe and efficient preliminary testing, providing robust data support for subsequent physical validation and significantly lowers the overall testing cycle.