Internal short circuit (ISC) is considered as one of the main causes of battery thermal runaway (TR), which has serious potential safety hazards for EVs. However, few electrothermal characteristics bring challenges for battery management system to diagnose ISC in the early stage, especially the accurate quantitative diagnosis. And most of the existing diagnostic methods horizontally compare the characteristics of different batteries, ignoring the impact of battery consistency. Therefore, a novel quantitative ISC diagnosis method based on analyzing the variation of the charging electric quantity in a fixed voltage window of each cell before and after ISC occurrence is proposed, which successfully applies to the diagnosis of traditional laboratory constant-current charging condition and real vehicle step-current charging condition. The fixed voltage window is optimized by the genetic algorithm, and the application conditions of the optimal results are further discussed. The electric quantity growth rate between adjacent cycles is put forward to identify the ISC cell in combination with set threshold value. The proposed method can effectively avoid the consistency issue by longitudinal self-comparison, and the quantitative estimation accuracy is within 6 %.