The control of the charging and discharging process of the power lithium battery is the key to its efficient operation. However, there are many complex electrochemical reaction processes, so the power lithium battery is difficult to model. In order to study a more suitable modeling method for the charging and discharging process of lithium batteries, this paper uses the model identification method of ARMAX and Hammerstein-Wiener to carry out a data-driven modeling of the relationship between the input current and the open-circuit voltage of the power supply during the charging and discharging process of lithium batteries. Through parameter identification, this paper compares the ability of two lithium battery system identification models to predict the battery voltage response. Through the identification of parameters, this paper compares the ability of two lithium battery system identification models to predict battery voltage response. The results show that this process has certain nonlinear characteristics, the predictive ability of the linear model is insufficient, and the nonlinear model is more suitable for predicting the dynamic behavior of this process, which provides a reference for future health control.