With the development of aviation electromechanical system, the detection and life prediction of battery become one of the important work of ground crew. In view of the difficulty of lithium battery life prediction under the condition of small sample, the advantages and disadvantages of GM (1:1) model, grey Verhulst model and neural network model in solving the problem are studied. Through the analysis of existing models, a grey Verhulst neural network model is proposed, which makes up for the defect of low prediction accuracy in the medium and long term of grey model, and reduces the influence of neural network on the sample size requirement. Taking the lithium battery of a certain type of aviation equipment as an example.The results show that the prediction accuracy of grey Verhulst neural network model is 0.7%, which is far lower than that of other models. It shows that the model has high accuracy, which proves the feasibility and effectiveness of the proposed method.