Academic self-concept refers to a set of stable perception, experience and evaluation that an individual has about his or her academic development in academic achievement situation ( Guo cheng, He Xiaoyan , 2006). It is an important index to predict one’s academic achievement and a sensitive factor to predict one’s personality wholeness, thus, an effective education should promote the formation and development of one’s positive academic self concept. However, present research on adolescents’ academic self-concept is weak and the participants concerned are Chinese Han majority from economic development area. The studies on ethnic minority from southwest China are rare. Hence, the current study was conducted to assess the characteristic of academic self-concept among four minorities so as to provide evidence for educating and cultivating students’ academic self-concept in southwest China. The study made use of the Academic Self-Concept Questionnaire, which was developed by professor Guo cheng and was later revised by Zhao Xiaoyun, as a measure of students’ academic self-concept. A total of 452 high school students participated in this study; all of them are ethnic minority of Tujia, Miao, Dong and Tibetan in Southwest China. Results showed that: (1) the development level of minorities’ academic self concept was in the middle, and no significant difference among the four ethic group was found. But in the academic achievement value dimension, the mean score in Tibetan students was significant lower than that in Tujia, Miao, Dong students. (2) In total academic self-concept, gender differences were not significant among four minorities. However, In the academic ability perception dimension, gender differences closed to significant (p = 0.059, p = 0.051, p = 0.084), the boys in Tujia, Dong and Tibetan ethnic minority had a higher mean score than girls, with a significant difference. In the academic behavior self-control dimension, boys had a higher mean score than girls in Tujia ethnic minority, with a significant difference（p < 0.05）. In the academic achievement value dimension, girls had a higher mean score than boys in Miao ethnic minority, with a significant difference（p < 0.05）.（3）Grade differences were significant only in Tibetan students（p < 0.05）, but in the academic achievement value dimension，students of grade three had a lower mean score than students of grade one（p < 0.05） and grade two（p < 0.01）in Miao ethnic minority , with a significant difference（p < 0.05）. In the academic ability perception dimension, students of grade one had a higher mean score than students of grade two（p < 0.01）in Dong ethnic minority. In the academic emotion experience and academic ability perception dimension, students of grade two had a higher mean score than students of grade one and three in Tibetan students. (4) urban-rural difference was significant in Tujia students and Dong students; students from rural areas had a higher mean score than students from urban areas, with a significant difference. However, no significant urban-rural difference was found between Miao and Tibetan students. On the basis of the above results, this study provide suggestions concerning education as follows: (1) great attention should be attached in the cultivation of high school students’ positive academic self concept in ethnic minority arrears in southwest China; (2) education department concerned should strength pertinence on ethnic difference when implement education on ethnic minority.