To study the disagreement between self-reported suicidal ideation (SR-SI) and clinician-ascertained suicidal ideation (CA-SI) and its correlation with depression and anxiety severity in patients with major depressive disorder (MDD) or bipolar disorder (BPD).Routine clinical outpatients were diagnosed with the MINI-STEP-BD version. SR-SI was extracted from the 16 Item Quick Inventory of Depression Symptomatology Self-Report (QIDS-SR-16) item 12. CA-SI was extracted from a modified Suicide Assessment module of the MINI. Depression and anxiety severity were measured with the QIDS-SR-16 and Zung Self-Rating Anxiety Scale. Chi-square, Fisher exact, and bivariate linear logistic regression were used for analyses.Of 103 patients with MDD, 5.8% endorsed any CA-SI and 22.4% endorsed any SR-SI. Of the 147 patients with BPD, 18.4% endorsed any CA-SI and 35.9% endorsed any SR-SI. The agreement between any SR-SI and any CA-SI was 83.5% for MDD and 83.1% for BPD, with weighted Kappa of 0.30 and 0.43, respectively. QIDS-SR-16 score, female gender, and ≥4 year college education were associated with increased risk for disagreement, 15.44 ± 4.52 versus 18.39 ± 3.49 points (p = 0.0026), 67% versus 46% (p = 0.0783), and 61% versus 29% (p = 0.0096). The disagreement was positively correlated to depression severity in both MDD and BPD with a correlation coefficient R 2 = 0.40 and 0.79, respectively, but was only positively correlated to anxiety severity in BPD with a R 2 = 0.46.Self-reported questionnaire was more likely to reveal higher frequency and severity of SI than clinician-ascertained, suggesting that a combination of self-reported and clinical-ascertained suicidal risk assessment with measuring depression and anxiety severity may be necessary for suicide prevention.