Abstract: When the electric vehicle is in different driving conditions, the SOC value can’t be accurately estimated by applying the Unscented Kalman Filter algorithm with fixed sampling rate (FSR-UKF). To overcome this shortcoming, the variable sampling rate acquisition method was combined with the traditional UKF algorithm according to the characteristics of the battery current in different driving states of electric vehicles. According to the equivalent circuit model of the controlled voltage source battery and its parameter identification results, this study applied the VSR-UKF algorithm to estimate SOC of battery and designed the estimation flow chart. Finally, the simulation verification experiment was designed. The simulation results show that the maximum relative error of the battery terminal voltage is less than 1% and the maximum SOC estimation relative error is less than 5%. So the correctness and accuracy of the study are verified.