To provide accurate and reliable prediction of the state of health of lithium batteries, a prediction model based on the African Vulture Optimization Algorithm fused with the Aquila Optimizer to optimize the BP neural network is proposed. Through the analysis of voltage, current and temperature data during the battery charging process, the correlation between the health factors and the SOH of the battery is verified based on grey correlation analysis, and four health factors are identified as inputs to the model, which are combined with the BP neural network model based on AO-AVOA optimization to achieve a more accurate SOH prediction. The proposed model is compared with other optimization models for SOH prediction of lithium batteries, and the results show that the average absolute error of less than 0.008 9, the root mean square error below 0.011 2, and the average absolute percentage error under 1.451 2%. This model is characterized by high accuracy and robust generalization capabilities, making it highly effective for SOH prediction in lithium batteries.