The prevalence of daily smoking has been estimated at the national and regional levels for nearly twenty years using data from the Santé publique France Barometer. At the departmental level, due to the insufficient sample size in the various editions of this survey, no results have ever been published. In this study, to provide the first departmental estimates of the prevalence of daily smoking, the Fay-Herriot (FH) small-area estimation method was used and compared to the direct method. Data on daily smoking come from the 2021 Santé publique France Barometer. These data covered 22,625 individuals aged 18 to 75 in mainland France. Sociodemographic data (age, sex, educational attainment, socio-professional category, household type, labor force participation rate, median income, etc.) from the 2019 population census were also used as auxiliary variables to predict daily smoking in the FH model. The departmental prevalences of daily smoking estimated by the direct method range from 12.6% to 41.4%. They range from 17.7% to 32.5% for the FH model. The precision of the estimated prevalence ranges from 7.4% to 53.1% for the direct method and is less than 15% for the FH method, regardless of the department. The results of this study demonstrate the feasibility of producing departmental estimates of daily smoking using data from the 2021 Santé publique France Public Health Barometer, which was initially intended to generate estimates at the national and regional levels. However, the selection of auxiliary variables that are well correlated with the variable of interest is essential to reduce bias in the FH model estimates.