The key to accurate simulation the thermal characteristics of lithium-ion batteries is to develop a reliable estimation model of the battery calorific value, which is very useful to design the thermal management system of the battery pack. This paper proposes a semi-empirical thermal model with virtues of high computational efficiency and precision for thermal analysis of the lithium-ion battery. In this model, a generalized and simple equation is developed to calculate the heat generation rate of the battery on the basis of its thermogenesis mechanisms. Two experimental fitting equations are acquired to estimate the battery ohmic resistance and entropy change coefficient, and a general formula is derived for the polarization resistance based on an approximate model of lithium diffusion process. Then, the semiempirical thermal model is established for the battery by using the proposed computational method of the heat generation rate. The galvanostatic discharging and charging processes of a prismatic lithiumion battery are separately simulated at the different currents and temperatures, and the corresponding experiments are conducted for the model validation. The results show that the battery temperatures simulated by the proposed model are good agreement with the experimental data. According to the calculation, the mean absolute deviations of the simulated battery temperatures are 0.27 degrees C for the discharging processes and 0.17 degrees C for the charging process under the orthogonal conditions of the operation currents (0.3C, 1.0C and 2.0C) and the ambient temperatures (10 degrees C, 25 degrees C and 35 degrees C). Compared with the theoretically based model, the proposed model exhibits a higher computational accuracy in the most operating regions and improves calculation speed by more than 4.0 times at the same grid density and time step setups. Consequently, the semi-empirical model proposed in this work is well-suited to pre-assess thermal characteristics and thermal management system schemes of the largescale battery pack due to its superiorities of easy implementation, high computational efficiency and high precision. (C) 2019 Elsevier Ltd. All rights reserved.