Accurate estimation of the state of health (SOH) of lithium-ion batteries is the critical to ensure the safety of lithium-ion batteries. However, the existing methods for SOH estimation of lithium-ion batteries exist unsatisfactory evaluation accuracy. To solve this problem, this paper proposes a battery SOH estimation method based on the combination of temporal convolutional network (TCN) and bidirectional gated recurrent unit (BiGRU). Firstly, the health factor is extracted from the battery charging data and its correlation with the battery capacity is discussed. Then, the TCN model is used to process the long series dependent data and carry out feature extraction, and also a Dropout layer is added to the model to prevent overfitting and improve the generalization. Finally, the BiGRU model is used to model the historical data features and predict the data degradation trend. In addition, the BiGRU model is used to model the historical data characteristics and estimate the data degradation trend to achieve an accurate assessment of the SOH of lithium-ion batteries. The results show that the proposed method obtains the better average of coefficient of determination (0.990 4), absolute mean error (0.017 1), and root mean square error (0.022 3) than other comparative methods under four batteries.