The aim of this study was to evaluate flossing as a diagnostic method for interproximal gingival bleeding in children. For this crossover study, 23 pre-schoolchildren presenting neither restorations nor approximal carious cavities and with at least 15% of gingival bleeding sites were selected. Examinations were performed at three different moments (3-4 days interval). Examinations comprised repeated measurements of two gingival indices with a 10-minute interval in the following sequences: the Ainamo & Bay Gingival Bleeding Index (GBI) followed by the Carter & Barnes flossing index (CBI); CBI followed by GBI; and GBI followed by GBI. Data analysis was performed only for the interproximal sites, considering the GBI as the gold-standard. Agreement between indices, sensitivity (SE), specificity (SP), positive (PPV) and negative predictive values (NPV) were estimated. Percentage agreements in sequences GBI-CBI, CBI-GBI and GBI-GBI were 70.3%, 76.4% and 84.5%, respectively.