In the automotive sector, integrated hardwaresoftware testing is crucial for ensuring software quality and reducing project costs. However, traditional methods face significant challenges due to the complex, implicit, and difficult-toextract nature of dependencies, which result in low efficiency and high costs in test case generation. To address this issue, this paper proposes an automated test case generation method that combines large language models (LLMs) with formal models to directly extract implicit dependencies from natural language requirement documents and generate integrated test cases. The proposed method leverages LLMs to model natural language requirements into structured flowcharts. It then analyzes these flowcharts to identify hidden dependencies and constructs sequential and temporal constraints for test scenarios. By converting dependencies into constraint equations and solving them using an SMT solver, the method generates input data for hardware devices, achieving fully automated test case generation. In a real-world case study with Beijing Automotive Industry Corporation (BAIC), the method successfully processed 300 functional requirements, extracting a total of 4731 implicit dependencies. The generated flowcharts achieved an accuracy rate of 81.24%, and the test case path coverage reached 82.67%. This significantly enhanced the level of automation in automotive integrated testing.