Dog rabies and human rabies transmitted by dogs can be prevented through adequate dog vaccination coverage. For surveillance purposes, this coverage must be calculated using accurate population size estimates to avoid misleading conclusions. We used data from a Brazilian national survey comprising 64,348 households, to calculate point estimates and confidence intervals of the size, household density, and rabies vaccination coverage of the owned‐dog population in Brazilian federative units. We also evaluated the precision of owned‐dog population size estimates, based on the extrapolation among different areas, of the mean number of dogs per household and the human/dog ratio. The estimated owned‐dog population size in Brazil was 52,198,324 (95% CI = 51,028,583–53,368,066) and the dog vaccination coverage was 80.09% (95% CI = 79.09%–81.09%). Both estimates had marked variation across Brazilian federative units and urban/rural strata. Only two of the 27 federative units had a confidence interval in rural vaccination coverage above 70% and six did not pass this threshold in their urban stratum. For the first time, we reported probabilistic estimates for an entire country and its main administrative areas. The estimated coverage for the country was high because the most populated federative unit also had high coverage. The mean number of dogs per household and the human/dog ratio were useful as survey estimates to characterize owned‐dog density. However, the simple extrapolation of these parameters resulted in estimates of owned‐dog population sizes with large errors (up to 254%) that must be interpreted with caution to avoid misleading conclusions. To evaluate the dog population size, we recommend the use of probabilistic sampling designs instead of simple human/dog ratio extrapolations, and the inclusion of animal‐related questions in censuses and national surveys to obtain reliable estimates to support improvements in animal and human health.