Abstract: Thermoelectric materials, as functional materials capable of directly converting thermal energy into electrical energy, are widely applied in multiple fields due to their environmental friendliness, low noise output, high operational reliability, and ease of miniaturization. Despite their promising prospects, the pursuit of high-performance materials faces challenges due to the high costs and time-consuming nature of experimental processes and simulation methods. In recent years, machine learning (ML), as a branch of artificial intelligence, has made a significant impact in the research and discovery of thermoelectric materials. This review examines the value of ML in predicting and optimizing the thermal and electrical transport properties of materials and explores its transformative role in accelerating the discovery and optimization of thermoelectric materials. Finally, it outlines the challenges and prospects of ML in thermoelectric research.