To improve the efficiency and safety, the State of Health (SOH) estimation of lithium-ion batteries is regarded as a key function of battery management system. In recent years, methods based on constant voltage charging data have been of great research interest. However, few of them explored the physical mechanism, causing the lack of generalization ability. To solve this problem, this paper analyzed the evolution mechanism of the constant voltage charging profile as the battery ages based on the electrochemical model. It is found that the stoichiometry is the key parameter to affect the constant voltage charging profile. Combined with the solid diffusivity, the two parameters are extracted from the electrochemical model as the health indicators (HIs) to estimate the battery SOH. The situation when faced with incomplete data is also discussed based on the particle swarm optimization algorithm. A public dataset is used to validate the proposed method. The result shows that the method has higher accuracy and is robust compared to the methods only based on data.