Objective: To explore risk factors for anorexia in children. Methods: A community-based 1:4 matched case-control study was conducted using questionnaire survey and vitamin A detection. Conditional logistic regression was used to estimate odds ratio. Results: Multivariate conditional Logistic regression analysis showed that Feeding without ration (OR=1.500,95%CI=1.071~2.101,P = 0.018),Partiality for a particular kind of food(OR=1.780,95%CI=1.267~2.501,P = 0.001),Be forced to feed by paterfamilias(OR=2.133,95%CI= 1.491~3.054,P < 0.001),Be ill with digestive diseases(OR=3.106,95%CI=1.955~4.933,P < 0.001),Be ill with respiratory infection in recent two weeks(OR=1.669,95%CI=1.112~2.513,P = 0.013),Feeding hardly just after birth(OR=1.946,95CI=1.284~2.950,P = 0.002) and Be lack of vitamin A (OR=3.222,95%CI=1.660~6.256,P = 0.001) are associated significantly with anorexia in children. Conclusion: Prevention and control measurements should be focused on the bad habits of children,nutritional knowledge of their parents and timely vitamin A rich foods intake to reduce the risk of anorexia.