Abstract: In order to make better use of the feature information obtained by all convolutional layers in convolutional neural networks（CNN）, a prediction model for the remaining life of lithium-ion batteries based on jump-connected multi-scale CNN is proposed. The model takes the health factor of the battery as input, uses the multi-scale CNN model based on jump connection, simultaneously extracts the local feature information and global feature information of different scales of the health factor of the lithium-ion battery, and fuses all the local feature information and global feature information through the information fusion module, and finally outputs the predicted value of the remaining life. Experimental results show that the proposed method can predict the remaining life of lithium-ion batteries more accurately. Compared with the classical CNN method, Bi-LSTM method, EMD-LSTM method and VMD-GRU method, the root means square error(E_RMSE) is reduced by 75.7%, 78.3%, 83.8% and 77.8%, respectively. Mean absolute error(E_MAE) decreased by 80.7%, 80.9%, 86.8%, 82.3%, and mean absolute percentage error(E_MAPE) decreased by 81.0%, 82.2%, 87.0% and 83.1%, respectively. The model determination coefficient（R~2） increased by 17.4%, 23.2%, 44.5% and 25.8%, respectively.