<p>Accurate estimation of the state of health (SOH) of lithium-ion batteries is crucial for ensuring their safety and usage. This paper proposes a lithium-ion battery state of health (SOH) estimation method using multi-feature fusion and the Swin Transformer model. Key health factors (HFs) related to capacity degradation are extracted from charge and discharge curves, and data preprocessing is performed using the Isolation Forest algorithm and different interpolation methods. The CEEMDAN method is employed to extract residual components that reflect battery degradation. The effectiveness of these health factors and residual components is verified through Pearson and Spearman correlation analysis, and key features are selected to construct a multi-feature fusion dataset. The paper also innovatively combines 1D CNN with 1D Swin Transformer to build a 1D CNN-Swin Transformer hybrid model, which fully integrates the local perception ability of convolutional layers with the Swin Transformer’s advantage in modeling long-range dependencies. The Swin Transformer reduces computational complexity through its shifted window design, enhancing computational efficiency while maintaining model performance. The proposed method is tested on NASA and CALCE datasets, showing significant improvements. On the NASA dataset, the <i>RMSE</i> metric effectively decreases by 11.83 to 32.14%, compared to LSTM, and on the CALCE dataset, <i>RMSE</i> metric effectively decreases by 40.64 to 58.76%.</p>