The prevalence of overweight and obesity reached 19.7% in 12-year-old French children in the year 2005. Recently, nationwide programs have been broadly implemented in France to reduce the overconsumption of sugars, salt and fat. The aims of this study were to assess the distribution of body mass index (BMI) and D(3+4)MFT index in a sample of 12-year-old French children, and to compare several regression models in order to analyze the association between these two indices. A cross-sectional study was conducted in Montpellier, France, and the height, weight, D(3+4)MFT, sugar and soft drink consumption were recorded in a randomly selected sample of 835 schoolchildren. In order to analyze the association between BMI and DMFT, four models of regression were tested: logistic, Poisson, zero-inflated Poisson (ZIP) and zero-inflated negative binomial (ZINB). The mean BMI was 18.9 for the whole sample and the corresponding DMFT value was 1.47. The caries prevalence was 51.7%. The best fitted models for testing the association between BMI and DMFT were ZIP and ZINB models. They showed a significant association between DMFT and sugar consumption, but not with BMI. As a result of the best fitted models (ZIP and ZINB), where BMI was not statistically associated with DMFT, we conclude, within the limits of a cross-sectional survey, that there is no association between these two variables