Abstract: Lightweight high-entropy alloys (LHEAs) are a new class of materials that exhibit lower density and superior specific strength compared to traditional alloys. They hold extensive potential for applications in advanced industries such as aerospace, automotive, and electronics. However, the current challenge lies in balancing the strength and ductility of LHEAs, as well as improving the manufacturing processes to reduce production costs and facilitate large-scale adoption. Consequently, the design and development of new LHEAs with low density and high strength have garnered significant attention. The precise and efficient application of machine learning methods in material design is observed by leveraging research findings from scholars both domestically and internationally. It elaborates on the use of machine learning to expedite the design of LHEAs at a low cost, including the implementation of high-throughput experiments and active learning strategies to continuously mine, clean, iterate, and optimize data. Additionally, it discusses the enhancement of model prediction accuracy and performance through feature engineering, the exploration of relationships between label values and feature variables using algorithms, and a deep analysis of the underlying reasons for the performance enhancement of LHEAs. The paper concludes with a summary and outlook on recent advancements in predicting the phase structure and mechanical properties of lightweight LHEAs using machine learning.