The accurate estimation of Electric Vehicle (EV) battery State of Health (SOH) is a critical component for ensuring vehicle reliability, optimizing performance, and assessing residual value. However, real-world applications are often constrained by the limited availability of comprehensive operational data, as direct sensor feeds are typically inaccessible to end-users and thirdparty systems. This survey provides a comprehensive review and comparative analysis of machine learning (ML) methodolo- gies designed to overcome this data-gap challenge. We investigate a spectrum of regression-based approaches, from traditional models like Linear Regression and Support Vector Regressors (SVR) to ensemble methods such as Random Forest and Gradient Boosting, and neural networks. A central focus of this review is the ”multi-stage prediction pipeline,” an architectural pattern designed to infer high-value SOH metrics from minimal, userpro- vided inputs. This pipeline operates in two distinct stages: (1) It first leverages easily obtainable data points, specifically Charging Duration, Total KM Traveled, and Battery Type, to model and predict a set of unobserved, intermediate operational parameters, including SOC , Battery Temp (°C), Ambient Temp (°C), and Charging Cycles. (2) These inferred features are then combined with the original user inputs to form a complete feature set, which is used to train a second set of models for predicting the final SOH indicators: Efficiency and Degradation Rate . By synthesizing the performance of various algorithms (evaluated using R² and MAE) within this two-stage framework, this paper identifies and validates this ”feature-inference” architecture as a robust and practical solution. It effectively bridges the gap between data in- tensive laboratory models and the exigent demands of real-world battery diagnostics, paving the way for scalable and accessible SOH estimation tools. Index Terms Electric Vehicles, Battery State of Health (SOH), Machine Learning, SOH estimation, Degradation prediction, Battery diagnostics, Feature inference, Multi-stage prediction, Regression analysis, Ensemble methods, Random Forest, Gradient Boosting.

^

[1]S. Pourbunthidkul, N. Pahaisuk, P. Laon, et al., “An enhanced cascaded deep learning framework for multi-cell voltage forecasting and state of charge estimation in electric vehicle batteries using LSTM networks,” Sensors, vol. 25, no. 12, p. 3788, 2025. [2]J. Gong, B. Xu, F. Chen, and G. Zhou, “Predictive modeling for electric vehicle battery state of health: A comprehensive literature review,” Energies, vol. 18, no. 2, p. 337, 2025. [3]R. Mamidi, D. Obulesu, K. B. Prajna, et al., “Enhancing battery health in electric vehicles: AI-enhanced BMS for accurate SoC, SoH, and fault diagnosis,” Metallurgical and Materials Engineering, 4th ed., 2025. [4]M. Ahwiadi and W. Wang, “Battery health monitoring and remaining useful life prediction techniques: A review of technologies,” Batteries, vol. 11, no. 1, p. 31, 2025. [5]N. Vasavi, A. A. Reddy, K. P. Chandra, et al., “Predicting EV battery lifespan using machine learning,” Int. J. Comput. Learn. Intell., vol. 4, no. 4, pp. 619–632, 2025. [6]M. Cavus and M. Bell, “Enabling smart grid resilience with deep learning-based battery health prediction in EV fleets,” Batteries, vol. 11, no. 8, p. 283, 2025. [7]S. Gu, K. Qian, and Y. Yang, “Optimization of electric vehicle charging and discharging strategies considering battery health state: A safe reinforcement learning approach,” World Electr. Veh. J., vol. 16, no. 5, p. 286, 2025. [8]Y. Wang, H. He, J. Zhang, et al., “EVBattery: A large- scale electric vehicle dataset for battery health and capacity estimation,” arXiv preprint arXiv:2201.12358v3, 2022. [9]K. R. Lin, A. L. S. Filipowicz, J. Li, and D. A. Shamma, “SOH illusion: Misunderstandings of EV battery state of health and methods to promote understanding,” 2023, pp. 3744333–3747828. [10]H. R. Hasan, K. Salah, A. Mayyas, et al., “Using composable NFTs and blockchain for the creation of EV battery digital passports with sustainability and traceability features,” Sustain. Futures, vol. 10, p. 100847, 2025. [11]A. Pohlmann, M. Popowicz, J.-P. Scho¨ggl, and R. J. Baumgartner, “Digital product passports for electric vehicle batteries: Stakeholder requirements for sustainability and circularity,” Cleaner Prod. Lett., vol. 8, p. 100090, 2025. [12]A. Ali, M. Al Bahrani, S. Ahmed, et al., “Sustainable recycling of end-of- life electric vehicle batteries: EV battery recycling frameworks in China and the USA,” Recycling, vol. 10, no. 2, p. 68, 2025. [13]M. Popowicz, A. Pohlmann, J. P. Scho¨ggl, and R. J. Baumgartner, “Digital product passports as information providers for consumers: The case of digital battery passports,” Bus. Strat. Environ., 2025. [14]Q. Li and J. Zhou, “A comparative analysis of extreme gradient boosting, decision tree, support vector machines, and random forest algorithm in data analysis,” Informatica, vol. 49, pp. 127–134, 2025. T. Champahom, C. Banyong, T. Janhuaton, et al., “Deep learning vs. gradient boosting: Optimizing transport energy forecasts in Thailand through LSTM and XGBoost,” Energies, vol. 18, no. 7, p. 1685, 2025.