The purpose of the present study was to develop a Chinese translation of the Observer Alexithymia Scale (OAS-C) and evaluate its reliability and factorial validity. The original English-version of the Observer Alexithymia Scale (OAS) was translated into Chinese and given to 468 Chinese undergraduate students. Students were asked to rate a person (other than themselves) whom they knew well (e.g., a parent, sibling, another relative, or friend). We evaluated internal consistency, test-retest and inter-rater reliability, and factorial validity. Average OAS-C scores were slightly higher than, but comparable to, OAS scores in the normative samples (English-speaking/nonclinical). The OAS-C showed adequate internal consistency (Cronbach's alpha coefficient was 0.84, and the mean inter-item correlation coefficient was 0.14), good stability (test-retest reliability with a 2-week interval was 0.90), and inter-rater reliability (intra-class correlation coefficient was 0.78). Moreover, the OAS five-factor model (Distant, Uninsightful, Somatizing, Humorless, and Rigid) was confirmed: incremental fit index=0.905, comparative fit index=0.904, and root mean square error of approximation=0.086; each represented an adequate model fit. The OAS-C appears to be a reliable and valid observer-rated alexithymia measure. We recommend that researchers collect both self- and observer-rated alexithymia data and, when possible, obtain observer reports from more than one person.