Background Eating behaviours have been suggested relating to obesity development. The Children's Eating Behaviour Questionnaire (CEBQ) is a parent-report measure constructed to assess multiple dimensions of eating behavior for children. This study aimed to test the validity of the Chinese version of Children's Eating Behaviour Questionnaire (CEBQ) in Chinese children aged 12-18 months. We examined factor structure and the reliability of the Chinese version of the CEBQ, the associations between children's eating behaviours and children's weight (BMI SDS) were assessed. Methods 219 questionnaires were filled out by the caregivers, approached in community health care centers in two cities in China. BMI of each child was calculated and converted to BMI SDS. Factor validation (Principal Component Analysis, exploratory factor analysis) on all CEBQ items was performed and gender difference in eating behaviours was examined. Correlations between eating behaviours and the child's BMI SDS were analyzed by linear regression analysis controlling for gender, parental combined weight, and education. Results The factor analysis revealed a seven-factor solution, with factor 'food responsiveness' (FR) split into two. 'Satiety responsiveness' (SR) and 'Enjoyment of food' (EF) factors were not detected. Interestingly, boys scored higher than girls in the FR scales, whereas girls had a higher score in 'food fussiness' (FF) scale. Conclusions We conclude that although a valuable psychometric instrument, CEBQ might be affected by age and cultural differences. Therefore, adjusting it in order to fit the Chinese population was suggested. We did not find an association between eating behaviours and children's BMI SDS, when it was controlled for gender and parental weight.