Abstract: As the energy power of electric vehicles, batteries are the most critical component in the performance and safe operation of electric vehicles, whose state of health (SOH) is of great significance to improve the safety and availability of new energy electronic equipment. Aiming at safe operation of lithium-ion batteries, this paper proposes a health state prediction algorithm based on feature fusion. The framework combines the health features of electrochemical impedance spectroscopy (EIS) and incremental capacity analysis (ICA), uses convolutional neural network (CNN) and the improved long short term memory (LSTM) network to establish the mapping relationship between features and state of health, and uses quantum particle swarm optimization (QPSO) algorithm to optimize the hyperparameters of the hybrid network structure. Finally, the NASA PCoE dataset is used to verify the accuracy and reliability of the method.