Heat generation during cycling of Lithium-ion batteries may have negative effects on the efficiency, safety, performance and lifetime of battery cell due to overheating. Therefore, temperature distribution should be controlled by the battery thermal management techniques. To improve accuracy of temperature control methods, the value of heat generation should precisely be calculated. However, detail modeling of phenomena, inside all cells of a battery pack, is a time consuming process, impracticable for control purposes. The prime objective of this paper is to propose a developed analysis viewpoint of heat generation phenomenon at pack scale by presenting a relation derived from the cell scale analysis without missing the accuracy along with vanishing computational cost. In order to increase the accuracy, firstly, the thermal modeling part at cell scale is improved to consider the heat of mixing and enthalpy heating terms as significant heat sources continue to be generated after cutoff current. Then, based on output data of the corrected code, a relation for heat generation source term is estimated by neural network method. Finally, this relation is used as a source term in pack scale equations. The results demonstrated the importance of considering heat of mixing and enthalpy heating in the pack scale modeling in a wide range of cooling air flow rates. This method is generic and vanishes the computational cost of transferring information from cell scale to the pack one. (C) 2017 Elsevier Ltd.