Aiming at the problem that there are many difficulties in accurately estimating the state of charge（SOC）value of lithium batteries

and at present

unscented Kalman filter（UKF）algorithm has the problem of poor stability and low precision

an improved UKF algorithm is proposed

which incorporates spherical point selection rules to reduce the impact of noise on simulation

and achieves accurate estimation of SOC of lithium batteries under non-linear conditions.The simulation results show that the error of the improved UKF in estimating SOC of lithium battery is less than 1 %

while that of the extended Kalman filter（EKF）and the standard UKF are 4 % and 2 %

respectively.Based on Thevinin battery model

SOC of lithium battery is estimated.The estimation on Li battery SOC based on Thevinin battery model is carried out.The precision of improved UKF algorithm is obviously higher than EKF algorithm and standard UKF algorithm.