Abstract: Modeling studies of lithium-ion batteries(LIBs) have been widely employed for performance analysis and design optimization due to their low resource consumption, short experimental cycles, and satisfactory accuracy. Mainstream modeling approaches are reviewed from both electrical and thermal perspectives and the coupling mechanism of electro-thermal models is explained. Electrical models can be categorized into electrochemical mechanistic models, equivalent circuit models and data-driven models according to the depth of physical interpretation. Electrochemical mechanistic models provide high accuracy and strong interpretability, but involve challenging parameter identification and high numerical complexity; equivalent circuit models feature simple structures and are well suited for online implementation; data-driven models depend on the coverage of training data and are effective for fast mapping under nonlinear and complex operating conditions. Thermal modeling consists of heat generation models and heat transfer models. The heat generation models are typically formulated based on the BERNARDI equation, while heat transfer models can be divided into lumped models, multi-dimensional models and equivalent thermal network models, which represent different trade-offs among temperature resolution, computational burden, and engineering deploy ability. On this basis, the electro-thermal coupling mechanism is clarified: the temperature dependence of electrical parameters and the impact of these parameters on heat generation and temperature. Finally, for different combinations of electrical and thermal models, specific coupling implementations and representative cases are presented, covering configurations ranging from high-fidelity electrochemical-multi-dimensional heat transfer models to high-efficiency equivalent-circuit-thermal network models. This review provides references for understanding LIBs modeling strategies and for selecting appropriate models in practice.