Background: This study examines the psychometric properties of the Mental Retardation Attitude Inventory-Revised (MRAI-R; Antonak & Harth) in Chinese college students. The research questions included: (1) Is the MRAI-R a reliable measure for Chinese college student? (2) Is the MRAI-R related to familiarity with people with intellectual disabilities (PWID) and professional training? and (3) Does the four-factor model of the MRAI-R fit the Chinese college student data? Method: Five hundred and thirty-four college students from China participated in the study. Descriptive, Pearson product-moment correlation, Cronbach alpha reliability coefficient and factor analyses were used to examine the reliability and validity of the MRAI-R in this Chinese college student sample. Results: The alpha coefficient reliability of the total scale of the MRAI-R was 0.78. The alpha coefficient reliabilities for the integration-segregation (INSE) subscale, the social distance (SDIS) subscale, the private rights (PRRT) subscale and the subtle derogatory beliefs (SUDB) subscale were 0.50 (INSE), 0.78 (SDIS), 0.50 (PRRT) and 0.21 (SUDB) respectively. Correlation analyses indicated that familiarity with PWID and professional training had a weak correlation with attitudes towards PWID. A confirmatory factor analysis revealed that the standardized factor loadings and goodness-of-fit indices were inadequate (the chi-square/degrees of freedom ratio = 3.24; the goodness-of-fit index = 0.85; the comparative fit index = 0.69; the RMSEA = 0.07). A principal axis factoring analysis (PAF) with oblique rotation identified three factors with 25% of the variance accounted for the sample. However, the results of this PAF were factorial complex and not interpretable in that many items of the scale had double loadings on more than one factor. Conclusions: The factor structure of the MRAI-R in this sale of Chinese college students did not replicate the structure found in American adults. Although the SDIS subscale of the MRAI-R appeared to be a reliable instrument among the four subscales of the MRAI-R, the reliabilities of the INSE, PRRT and SUDB subscales were low. Continuing investigations of the utility of the MRAI-R in Chinese culture is needed. Attitude instruments based on Chinese culture may be developed.