Abstract The well-known own-race effect (ORE) in facial recognition is regarded as better recognition for faces of one’s own race than others. Lots of researches focus on the relation between own-race effect and inter-group contact by comparing the white race faces to others, and generate inconsistent results. On the ground of the above reasoning, the current study is to explore whether own-race effect could be extended to the individuals who are racially homogeneous but from different ethnic groups. The influence of quantity and quality of inter-group contact on the degree of own-race effect has also been discussed. On the basis of population distribution and patterns of inter--group contact, 99 Uygur college students from different regions were therefore recruited as the participants. First, they were requested to recognize the Uighur and Han faces by using “learning-recognition” paradigm. Face stimuli were 40 female face photographs (20 Han and 20 Uygur) without outer face features. During study, the participants looked at the series of 20 faces (10 Han and 10 Uygur) presented randomly on the screen. Each face was shown for 5 sec with an interval of 1 sec. The participants were requested to remember the faces. At test, all 40 faces were presented randomly. The participants were asked to judge whether the faces were new or old, and the computer recorded the reflex index of each participant. Then they were asked to answer Social Contact Questionnaire and Individuating Experience Questionnaire (Walker & Hewstone, 2006a) to assess self-reported quantity and quality of contact. The results of repeated measures ANOVAS show a generalized own-race effect for the Uygurs. The dependent variables, such as accuracy rates, reaction times, hit rates, discrimination accuracy and response criterion, demonstrate significant own-race advantage of the Uygurs. The Uygurs with minimum interracial contact have the lowest accuracy rates but the highest false alarm rates of Han faces, while the Uygurs with maximum contact have no differences on Uighur and Han faces in accuracy rates and discrimination score. The size of own-race effect in recognition memory for each participant was also analyzed, which was calculated by subtracting other-race accuracy rates from own-race accuracy rates (Hancock & Rhodes, 2008). The results of regression analysis show that both quantity and quality of the contact can reduce the size of own-race effect. But dominance analysis shows that compared with the quantity of the contact individuating experience has greater impact on the size of own-race effect (71.86% vs 28.14%). It is also found that only 29.5% of the variance in the own-race effect is attributed to the quantity and quality of contact, which demonstrates other social variables may influence the size of own-race effect. On the basis of these findings, it concludes that own-race effect can be found not only between different races, but also between different nationalities within the same race. In addition, the current study suggests that individuating other-race experience is more effective than quantity of contact to influence the magnitude of own-race effect.