Abstract: Due to the state of health (SOH) prediction of pure electric vehicle power battery relates to many non-linear factors and complicated algorithms, it is difficult to accomplish in singlechip platform. In order to overcome the difficulty, a new method was proposed. Firstly, a method for counting the accumulative charging cycles was used to calculate the number of battery use cycles. Then the nonlinear relationship between SOH and the number of cycles, the variation of internal resistance and the value of voltage drop were transformed into a discrete two-dimensional datasheet. According to the use conditions, the SOH values under different estimation methods could be obtained by using the binary look-up table method. Secondly, taking the number of cycles, voltage drop and internal resistance variation as input and the weight of corresponding SOH as output, a SOH dynamic prediction model was established based on T-S fuzzy control. According to the weights and boundary conditions, the SOH could be calculated. The simulation results show that the proposed method has a maximum prediction error of 4.3% and a response time of 55 ms, and the prediction effect is much better than that of the existing methods.