Background: Odor identification deficits occur in individuals with schizophrenia and their unaffected first-degree relatives, while deficits are less pronounced in individuals with bipolar disorder. We hypothesized that children at familial high-risk for schizophrenia (FHR-SZ) show odor identification deficits compared to population-based controls and that children at familial high-risk for bipolar disorder (FHR-BP) perform intermediate. Methods: Odor identification was assessed at age 7 in 184 children with FHR-SZ, 106 children with FHR-BP, and 186 population-based controls with the Brief Smell Identification Test. Dimensional and predefined categorical outcomes were used in the analyses. Potential relationships with psychopathological, cognitive, and home environmental variables were conducted using hierarchical and logistic multiple regression analyses. Results: ANOVA revealed no between-group differences in odor identification. Using the recommended cut-off (below 5), we found a significantly greater proportion of boys at FHR-SZ than population-based boys with an abnormal odor identification (p = .013). However, a supplementary analysis using a Danish-based cut-off (below 4) did not support this. All children showed significant, positive associations of odor identification with female gender, social responsiveness, and verbal working memory. Lower social responsiveness predicted abnormal odor identification in boys at FHR-SZ, only using the recommended cut-off. Conclusions: Odor identification efficacy and risk status appear independent in this early developmental phase. Using the recommended threshold, abnormal odor identification is more frequent in young boys at FHR-SZ than in population-based boys and is linked to lower social responsiveness. The validity of these results is questioned by non-significant differences in the rates when using an exploratory Danish-based threshold.