Aims. Whether or not cannabis use may increase or decrease the risk of type 2 diabetes is not clear. We analyzed the association between cannabis and subsequent type 2 diabetes and if a potential positive or reverse association persisted after controlling for potential confounders. Methods. In this population-based cohort study, 17,967 Swedish men and women (aged 18–84 years), who answered an extensive questionnaire in 2002 (including questions on cannabis use), were followed up for new cases of type 2 diabetes (n = 608) by questionnaire (in 2010) and in health registers during 2003–2011. Odds ratios (ORs) with 95% CIs were estimated in a multiple logistic regression analysis. Potential confounders included age, sex, BMI, physical inactivity, smoking, alcohol use, and occupational position. Results. The crude association showed that cannabis users had a reduced risk of type 2 diabetes OR = 0.68 (95% CIs: 0.47–0.99). However, this inverse association attenuated to OR = 0.94 (95% CIs: 0.63–1.39) after adjusting for age. Conclusions. The present study suggests that there is no association between cannabis use and subsequent type 2 diabetes after controlling for age. To make more robust conclusions prospective studies, with longer periods of follow-up and more detailed information about cannabis use, are needed.