Purpose: Current literature indicates that childhood obesity is associated with family structure type. For example, children living in grandparent-headed households are more likely to be obese compared to children living in two-parent headed households. However, little is known about the conditions that might interact with family structure to influence childrenâs obesity. Since family protective factors are known to directly influence childhood obesity, we examined whether two specific family protective factors might moderate the association between family structure type and childhood obesity. Methods: We utilized data from the 2017-2018 National Survey of Childrenâs Health (n=24,800). Multivariable logistic regressions were used to assess the moderating effect of âfamily resilienceâ and families eating meals (7 vs. < 7 days/week). We controlled for childâs sex, age, race/ethnicity, health insurance, number of adverse childhood experiences, and physical activity level. We tested the moderating effect of family resilience and family meals as dichotomous variables, on the association between family structure type and obesity. Results: In the fully adjusted models, we observed no significant moderating effect of family resilience on the association between family structure type and childhood obesity. There was a moderating effect of the family meal variable: In grandparent-headed households (compared to two-parent headed households), children were significantly less likely to be obese when families ate meals together 7 days/week (adjusted odds ratio [aOR] =0.95) vs. < 7 days/week (aOR = 2.55). In single-parent headed households, the association did not differ significantly by that variable (aOR =1.02 vs. 1.46). Conclusion: Our results provide evidence that eating meals together is one family protective factor that could be important considerations when developing family-based obesity interventions for children living with their grandparents. Disclosure M. Song: None. L. L. Hayman: None. N. Dieckmann: None. C. Musil: None.