Background and aims Lottery gambling participation tends to be higher among lower socio‐economic status (SES) individuals, but it is unclear how this relationship differs as a function of lottery type. We estimated how the relationship between SES and lottery gambling rates varies across different types of lottery gambling: fixed‐prize, progressive‐prize (jackpot) and instant‐win (scratch card) lottery tickets in a large Canadian city. Design Neighborhood‐level lottery purchase data obtained from the Ontario Lottery and Gaming Commission were analysed in conjunction with demographic data. Mixed‐effects regression was used to assess simultaneously how neighborhood‐level SES predicts per‐person lottery gambling rates across fixed‐prize, progressive‐prize lottery and instant‐win lotteries. Setting and participants Neighborhoods in Toronto, Ontario, Canada in the years 2012–15. Measurements Per‐capita sales in dollars (CAD) of fixed‐prize lottery, progressive‐prize lottery and instant‐win tickets in Toronto postal codes. SES was estimated as a composite of income, years of education and white‐collar employment. Findings Lower‐SES neighborhoods engaged in higher rates of lottery gambling overall β = −0.084, standard error (SE) = 0.24, P = 0.0007. The predictive effect of SES varied significantly by lottery type (fixed‐prize: β = −0.105, SE = 0.004, P < 0.0001, instant‐win: β = −0.054, SE = 0.004, P < 0.0001; relative to progressive‐prize). The predictive effect of SES was strongest for fixed‐prize lotteries and weakest for progressive‐prize lotteries, such that we did not observe a significant predictive effect of SES for progressive‐prize lotteries (β = −0.031, SE = 0.024, P = 0.198). Conclusions People in lower socio‐economic status neighborhoods in Toronto, Canada appear to engage in more lottery gambling than those in higher socio‐economic status neighborhoods, with the difference being largest for fixed prize lotteries followed by instant win lotteries, and no clear difference for progressive prize lotteries.