After years of public debate about same-sex marriage, the Australian Government put the issue to the electorate in the “Australian Marriage Law Postal Survey” in late 2017. The survey asked voters whether the law should be changed to allow same-sex couples to marry. Nationally, 61.6% of voters responded “Yes.” But there were marked variations by electoral division, with the proportion of “Yes” votes varying from 26.1% to 83.7%. The aim of this paper is to explore the geographical pattern of the percentage of voters responding “Yes” by federal electoral division and identify its correlates. Results of the survey by federal electoral division were obtained from the Australian Bureau of Statistics (ABS); other variables for electoral divisions were obtained from the ABS and the Australian Electoral Commission. Multiple linear regression and geographically weighted regression were employed to establish the relationship between the percentage of “Yes” responses at the electoral division level and the characteristics of electoral divisions’ populations. In 133 of 150 electoral divisions there was a majority of “Yes” responses. Strong predictor variables of the percentage “Yes” vote included the proportions of: the population describing themselves as having no religion, those with post-school educational qualifications, those with a birthplace in Oceania, Europe, or the Americas, and those who did not vote for conservative parties in the 2016 federal election. A marginally better fit was obtained with geographically weighted regression. In conclusion, the geographical pattern of responses in the Australian Marriage Law Postal Survey is closely associated with a small number of characteristics of an electoral division's population.