Background <p>Suicidal ideation (SI) is defined as death-centered thoughts preceding suicidal behaviors, and demands prompt identification. While depressed patients are disproportionately affected by SI, particularly when comorbid with anxiety, the key triggers driving its emergence remain poorly understood.</p> Methods <p>This study analyzed clinical data from 9,193 psychiatric outpatients, with in-depth assessments conducted on 281 depressed patients (mean age 41.9 ± 21.5; 73 males) who completed the Self-rating Depression Scale (SDS) and Self-rating Anxiety Scale (SAS). Of these, 85 patients (30.2%) reported SI. To characterize the interplay among symptoms, we applied a dual-network approach: a regularized partial correlation network quantified conditional associations between SDS/SAS symptom dimensions and SI, and a Bayesian network explored potential precursors to SI.</p> Results <p>In the regularized partial correlation network, SI was associated with two symptom dimensions: ‘anxiety and panic’ (edge weight = 0.14) and ‘depressed mood’ (edge weight = 0.16). ‘Anxiety and panic’ exhibited the highest strength centrality (z = 1.43), indicating its pivotal role in mediating the development of SI. Bayesian network analysis further identified ‘anxiety and panic’ as a key mediator in all indirect paths linking depressive symptoms to SI.</p> Conclusions <p>This dual-network approach highlights the utility of computational psychiatry in revealing the relationships among symptoms within a network, yielding clinically actionable information. Here, we uncovered the central role of anxiety within the symptom network underlying SI, offering insights relevant to early identification and suicide prevention of patients with comorbid depression–anxiety.</p>