To establish a graphic model of functional variables for depression according to the International Classification of Functioning, Disability, and Health (ICF) to explore the key target items and intervention paths, so as to provide new ideas and a basis for diagnosis and treatment of depression and recovery of psychosocial function. Methods: A total of 175 patients with depression were selected as research subjects. Patients were assessed by self-designed general condition questionnaire, patient health questionnaire (PHQ-9), Generalized Anxiety Disorder Scale (GAD-7), 24-item Hamilton Depression Scale (HAMD-24), and ICF Core Sets for depression. R software was used for modeling variable relationships and network analysis. Results: (1) A total of 62 ICF items were located in the main structure of depression, among which 27 items were located in the 4-coreness structure, 30 in the 3-coreness structure, and 5 in the 2-coreness structure. There was no significant correlation between the disability rate of each item and network parameters (P>0.05). (2) The results of frequency distributions showed that the range of the disorder rate of 27 items in the 4-coreness structure was basically consistent with the frequency distribution of the dysfunction rate of each item, and all items in the 4-coreness structure could be used as key intervention targets. (3) The results of risk correlation showed that items like optimism, thought functions, motivation, sensations associated with the digestive system, and sensation of pain had high risk correlation with multiple items, which can be used to guide the design of different intervention paths. Conclusion: The key target items and intervention paths can be explored by constructing the graphical model of ICF functional variables for depression, which will further guide the diagnosis, treatment, and recovery of psychosocial function in depression.