The electrical coupling of a synchronous condenser and its excitation system makes parameter identification difficult, and the analysis of the coupling characteristics between the identified parameters and observed variables is omitted. Therefore, a parameter classification and identification method based on sensitivity analysis is proposed. The key parameters of the synchronous condenser are analyzed at different time scales, and the types of parameters are classified based on the trajectory sensitivity information to establish the sets of coupling parameters. Considering the differences in response characteristics of the synchronous condenser and its excitation system in steady-state, sub-transient, and transient conditions, the parameter identification models are separately established at different time scales. The parameter identification is carried out according to the coupling parameter sets, and an improved snake optimization algorithm based on Tent chaotic mapping is adopted for analysis. The effectiveness and adaptability of the proposed method are verified by case study analysis.