Abstract Background Our purpose was to examine the reliability and validity of the Chinese version of the Patient Health Questionnaire-15 (PHQ-15) in Taiwan, and to explore its relation to somatoform disorders (DSM-IV) and to somatic symptom and related disorders (DSM-5). Methods We recruited 471 individuals, 151 with somatoform disorders and 200 with somatic symptom and related disorders. Subjects completed the Chinese version of the PHQ-15, Beck Depression Inventory-II (BDI-II), Beck Anxiety Inventory (BAI), and received a DSM-IV- and DSM-5-based diagnostic interview. We performed exploratory factor analysis and assessed test-retest reliability, internal consistency, and correlation with BDI-II/BAI to confirm reliability and validity, and carried out ROC curve analysis to determine suitability for evaluation or screening purposes. PHQ-15 scores were compared between patients with various DSM-IV psychiatric diagnoses (such as DSM-IV somatoform disorders, panic disorder, other anxiety/depressive disorders) or no DSM-IV diagnosis and patients with DSM-5 somatic symptom and related disorders or no DSM-5 diagnosis. Results The Chinese version identified cardiopulmonary, pain-fatigue, and gastrointestinal as major factors and had good reliability (0.803–0.930), internal consistency (0.637–0.861), and correlation coefficients with BDI-II/BAI (0.407–0.619, 0.536–0.721, respectively). The PHQ-15 scores were similar in patients with somatoform disorders and patients with panic disorder; higher in patients with somatoform disorders and panic disorder than in patients with other anxiety/depressive disorders; and significantly higher in patients with somatic symptom and related disorders than in patients without this diagnosis. The AUC of the PHQ-15 was 0.678 (cutoff 6/7) for screening somatoform disorders (DSM-IV) and 0.725 (cutoff 4/5) for screening somatic symptom and related disorders (DSM-5). Conclusions The Chinese version of the PHQ-15 is suitable for evaluating somatic symptom and related disorders. The preponderance of somatic symptom disorder in our sample, lack of evaluation of functional disorders, and recruitment solely from psychiatric clinics are possible limitations.