As a non-destructive method for characterizing battery dynamic behavior, electrochemical impedance spectroscopy (EIS) is becoming increasingly important to analyze electrode process kinetics, electric double layers, and diffusion in the study of battery performance. However, overlapping features of the EIS semicircle of commercial batteries over the lifetime bring ambiguities concerning their physicochemical significance spectra in the timescale distribution. In this work, we analyze the impedance of kinetic processes and corresponding time constants in three types of commercial batteries at different aging stages via the optimized distribution of relaxation time (DRT) investigation which extracts evolution details of different time scales that could not be distinguished. Firstly, informative aging tests for battery LiNi_0·8Co_0·1Mn_0·1O₂ (NCM), Li_xFePO₄ (LFP), and LiNi_xCo_yAl_zO₂ (NCA) during the whole life cycle were designed to simulate different electric vehicles working conditions, followed by periodical EIS measurements. And then, the capacity retention, as well as the deconvolution of EIS for discriminating the electrochemical mechanisms of commercial LiBs during aging were compared. Gaussian process and ridge regression are dedicated to complex superposed impedance spectra to DRT. Combining both experimental measurements and multi-peak analysis for DRT, the analysis determined the impedance distribution characteristics of batteries with different materials during cycling, which facilitates rapid identification of aging mechanisms and further life prediction. Therefore, the peak characteristic analysis method based on the DRT analysis can be employed to diagnose and predict the impact of battery performance.