A comprehensive review of machine learning-based state of health estimation for lithium-ion batteries: data, features, algorithms, and future challenges

Lithium-ion batteries are widely used in electric vehicles and energy storage systems, where reliable state of health (SOH) estimation is critical to ensure operational safety and lifecycle management. In recent years, machine learning (ML) has emerged as a powerful data-driven tool for battery SOH estimation due to its capability in learning complex degradation patterns from data. This review systematically examines the full ML-based SOH estimation workflow, beginning with an overview of measured, public, and synthetic datasets and common preprocessing techniques. The review then emphasizes feature engineering, analyzing the extraction of health indicators from voltage, current, temperature, incremental capacity (IC) curves, and advanced sensor data, as well as feature selection methods that improve model efficiency and robustness. On this basis, various ML algorithms are evaluated in terms of accuracy, generalization, practical deployment, and other key attributes. Finally, the review discusses ongoing challenges such as data scarcity, domain transferability, and physical consistency, and highlights future directions including physics-informed learning and hybrid data–model fusion. This work offers a comprehensive reference and forward-looking insight into the development of intelligent, reliable, and interpretable battery health estimation systems.