The residual life is one of important performance of lithium-ion battery. Before the life prediction, the SoH (State of Health) data of lithium-ion battery are necessary to be available. In order to improve the accuracy of SoH estimation, electrolyte dynamics is added to the single particle model of lithium-ion battery in this paper. Then, a novel Pade approximation and least squares method are employed to estimate the SoH of lithium-ion batteries. After that, the mapping particle filter is applied to forecast the battery life. MPF can greatly improve the diversity of particles and avoid the operation of resampling. This is the first time that the mapping particle filter has been used to forecast the residual life of lithium-ion batteries. Finally, the experimental data from National Aeronautics and Space Administration is used to prove that the mapping particle filter has a higher precision of prediction than the standard particle filter.