Objective Structural equation model (SEM) is a kind of modeling method on establishment,estimation and test of causal relationships.Currently,SEM is widely used in the fields of psychology,education and social science.There are many important latent variables (e.g.syndrome),which cannot be directly measured in traditional Chinese medicine (TCM) clinical data.The traditional one-order confirmatory factor analysis (CFA) method can only analyze the correlation between one-level latent variables and symptom indicators.We propose the two-order CFA modeling method based on TCM clinical data to analyze the correlation between two level syndromes and indicators,which can provide help for doctors to further understand disease progress quantitatively.Methods According to the characteristics of the TCM clinical data,we treated the TCM and deficiency-excessiveness differentiation syndrome as two order factors,and performed the latent variable modeling analysis on them and indicators.Results The results demonstrated the relationships of syndromes and deficiency-excessiveness differentiation syndromes,and the relationships of syndromes and different symptom indicators in non-small cell lung cancer.Conclusions There are certain inherent rules between the syndromes and the indicators.Our study has certain significance on the standardization of TCM syndrome.