Context: In a large-scale system such as the Android Operating System (OS), new features and updates are regularly added to the system; it is necessary to ensure the quality of these changes and not compromise the existing software. Objective: This paper presents a method for generating functional test cases for Android systems using the Large Language Model Meta AI (LLama) to reduce the manual test effort and avoid process bottlenecks. Method: Creation of prompts in LLM for test case generation using Android feature specification to design test cases to interact with Android platform. A feasibility study was performed to evaluate the test case writing quality based on clarity, objective, completeness, and maintenance, as well the test case complexity based on initial setup, number of steps, software components coverage, and abstraction level. Results: 12 Android features were evaluated, and the results obtained 57 test cases generated covering 100% of requirements. 19.3% of the test cases were considered well-written, and 63.2% had acceptable writing quality. 26.3% of the test cases had a medium complexity, and 73.7% had a low complexity. Conclusion: This method performed well in terms of writing quality for Android System. However, regarding complexity, the test cases had some non-existent initial setup and few interactions with components for the features. As lessons learned, the simpler the test suite the better test cases are generated, and this method needs effort to provide detailed instructions and UI information.