Life prediction for Lithium-ion batteries is an important means to master the decline tendency of power performance, and the life prediction methods of Lithium-ion batteries have become the research hotspot in the electronic system field of Prognostic and Health Management. Aiming at the life prediction of Lithium-ion battery, based on the data collected from the Lithium-ion battery ground test of NASA Ames center, the extended Kalman filter (EKF) algorithm is proposed and applied to the prediction process of Lithium-ion life, and it is modified by using the optimal Loess smoothing principle, which improves the stability and accuracy of prediction. The experimental results show that the proposed prediction method can be effectively used in the life prediction of Lithium-ion batteries, and has high practical value in engineering application..