Background: Boys use cannabis at a younger age and more frequently than girls. It has been suggested these sex differences might vary according to students' relationship to school. We explored whether the association between sex and adolescents' cannabis use varies among schools and according to students' school connectedness. Methods: The study population consisted of all students from 11 secondary schools in the greater Québec City area. The sample included 6185 respondents in years 1 to 5 at the secondary level (equivalent to grades 7-11). Study outcomes were monthly cannabis use and early cannabis use. Results: The association between sex and monthly cannabis use varied significantly among schools after controlling for students' main characteristics and school socioeconomic environment. We found a statistically significant modifying effect of school connectedness on the association between sex and monthly cannabis use. For early cannabis use, we found no modifying effect of school connectedness nor any association with sex. Conclusions: Measures to reduce adolescents' cannabis use could be better adapted to local context and more tailored to specific higher-risk groups. School connectedness is a protective factor for cannabis use, although this effect appears stronger for girls than boys.