In order to estimate influenza-associated excess mortality in southern Brazil, we applied Serfling regression models to monthly mortality data from 1980 to 2008 for pneumonia/influenza- and respiratory/circulatory-coded deaths for all ages and for those aged ⩾60 years. According to viral data, 73·5% of influenza viruses were detected between April and August in southern Brazil. There was no clear influenza season for northern Brazil. In southern Brazil, influenza-associated excess mortality was 1·4/100 000 for all ages and 9·2/100 000 person-years for persons aged ⩾60 years using underlying pneumonia/influenza-coded deaths and 10·0/100 000 for all ages and 86·6/100 000 person-years for persons aged ⩾60 years using underlying respiratory/circulatory-coded deaths. Influenza-associated excess mortality rates for southern Brazil are similar to those published for other countries. Our data support the need for continued influenza surveillance to guide vaccination campaigns to age groups most affected by this virus in Brazil.