A method for extracting the driving habit-specific features related to capacity degradation is proposed to address the issue of battery degradation variability in automotive lithium-ion batteries when the driving habits are different. The key factors affecting battery degradation under different driving habits are the different segmented acceleration processes and their proportions. A differential feature extraction method is designed. First, based on CLTC-P with a minimum time of 0.1 s and real vehicle test data, the driving habits are quantified by average speed and the proportion of the acceleration process, and the differences in the driving habits are analyzed. Then, the principal component analysis method is used to extract key influencing factors of the current, and the BI-KMEANS clustering method is employed to analyze pulse currents, combining the current differences under different driving habits to extract relevant features. Finally, the experiments are designed based on the differential features to explore their association with capacity degradation. The result shows that the different polarization times and intensities brought by different segmented acceleration processes are the main reasons for the variability in battery degradation when there are different driving habits.