<p>The remaining useful life (RUL) of lithium-ion batteries is a key parameter of battery management systems, and accurate prediction is an important guarantee for the stable and efficient operation of new energy vehicles. A fusion algorithm based on a multi-layer kernel extreme learning machine and genetic particle swarm optimization algorithm is constructed to study the prediction of lithium-ion battery RUL and improve the accuracy and stability of the single battery. The experimental verification results show that the RUL prediction accuracy of the fusion algorithm is significantly higher than other comparison algorithms in the long- and short-term prediction of single battery and cross-training validation, indicating that the fusion algorithm has high prediction accuracy and strong generalization performance in the RUL prediction, which provides a reference for the safety management and maintenance of lithium-ion battery in practical applications.</p>