To detect the health of lithium batteries and for battery fault diagnosis, and based on battery life cycle degradation data, increment capacity analysis and differential voltage analysis are used to extract features. The Pearson correlation coefficient is used to analyze the correlation of health factors. Health factors are input into an artificial neural network for battery state of health (SOH) prediction. From the battery capacity nonlinear degradation characteristics and its regeneration phenomenon, a dual exponential function is used to model it. A particle filter algorithm is used to estimate model parameters and achieve the probability density prediction of remaining useful life (RUL). The experimental results show that the proposed algorithm can achieve accurate prediction of SOH and uncertainty estimation of RUL. This work is supported by the General Project of National Natural Science Foundation of China (No. 62176227).