Abstract: Tuberculosis (TB) remains a serious threat to global public health. Traditional prevention and control approaches face bottlenecks in diagnosis, early screening, drug resistance monitoring, and primary healthcare capability building. The rapid development of artificial intelligence (AI) technology has provided a new approach for comprehensive TB prevention and control. This paper systematically summarizes the progress in application of AI in TB imaging diagnosis, sputum bacterium detection, drug resistance prediction, drug development, epidemiological surveillance and patient management, introduces the advantages of AI in improving diagnosis efficiency, reducing missed diagnosis and misdiagnosis rates, optimizing prevention and control protocol, and facilitating grassroots screening. and analyzes the challenges currently facing AI applications, including insufficient data standardization, poor model generalization, barriers to clinical translation, and lack of ethical supervision. Furthermore, it predicts the future development of AI-aid TB prevention and control from the perspectives of data system construction, technology integration and innovation, scenario-based deployment at the grassroots level, and improvement of the policy ecosystem for the purpose of providing reference for the intelligent system establishment and strategy development for TB prevention and treatment.