The Hospital Anxiety and Depression Scale (HADS) is a widely used screening tool for depression and anxiety in patients with medical illnesses. This study aimed to explore the psychometric properties of the HADS in screening for patients with major depression using item response theory method. A total of 460 patients with major depression completed the HADS. Rasch analyses were used to examine unidimensionality, item fit, local dependency, reliability, ordering of categories, differential item functioning (DIF) and targeting. The HADS showed a two-dimensional construct. All items fit the Rasch model. Three pairs of items showed minor but inconsiderate local dependency. Both subscales had acceptable reliability. None of the items displayed disordered categories or DIF. All items were well targeted, and participants with high and low levels of distress were less targeted than those with moderate levels of distress. Finally, a conversion table to transform the raw scores into interval measures was generated. The HADS demonstrated adequate psychometric properties in assessing depression and anxiety in patients with major depression. It was more appropriate for assessing moderate than high or low levels of distress. The conversion table can be used for more precise measurement. These results may pave the way for efficient and sensitive methods of analyzing depression symptom response in research and in clinical practice.