Given the promising prospects of retired lithium-ion batteries in second-life utilization, enhancing their consistency through a rational sorting process has become a pressing priority. Traditional capacity-based sorting methods have significant limitations as it takes high time costs and fails to provide internal dynamic information about the batteries. To address this, the present study introduces a novel approach by incorporating electrochemical impedance spectroscopy (EIS) into the sorting process. Firstly, principal component analysis (PCA) analysis is applied to extract the first principal component from the EIS data, which has a strong correlation with battery capacity. It serves as a key feature for assessing the residual value of retired batteries. Accurate estimation of battery capacity is then achieved using a simple linear equation: For retired nickel-cobalt-manganese (NCM) batteries, the mean absolute percentage error (MAPE) and root mean squared percentage error (RMSPE) for estimated capacity are 3.12% and 4.49%, respectively. While for retired lithium iron phosphate (LFP) batteries, the values are 4.47% and 6.16%. To characterize the differences in internal states among various batteries, distribution of relaxation time (DRT) is employed to analyze the electrochemical processes within the EIS. The similarity of the DRT curves is then calculated using the dynamic time programming (DTW) algorithm. Subsequently, the DRT curve similarity that indicates internal state is combined with external indicators, including estimated capacity and ohmic impedance, to serve as sorting factors. Finally, a two-stage sorting scheme is proposed: The density-based spatial clustering of applications with noise (DBSCAN) algorithm identifies abnormal batteries and performs preliminary sorting, followed by the greedy algorithm for further grouping based on the preliminary sorting clusters. The sorting results of the two retired batteries demonstrate that the above method can achieve comprehensively consistent sorting of the external characteristics and internal states, injecting new impetus into a more accurate and efficient sorting process.