ABSTRACT The objective of this study was to provide rates of cannabis use and dependence and risk factors, proposing a conditional path model for cannabis addiction. A subsample of adult participants from a Brazilian household survey was analyzed to estimate cannabis dependence. Estimation of prevalence rates and association between dependence and age of cannabis use initiation were performed. The conditional model was applied to investigate the mediation of depressive symptoms and alcohol consumption in the association of early cannabis use and dependence. Lifetime and last year cannabis use were 6.47% and 2.81%, respectively. Moderate and severe cannabis dependence were 0.74% and 0.18% in the population, and 26.47% and 6.38% among last year’s cannabis users. A Parallel Multiple Mediator Model revealed strong association between cannabis use initiation age and dependence, when depressive symptoms or alcohol consumption mediate this association. The proportion of cannabis users displaying dependence symptoms is elevated and it is associated with the age of cannabis use initiation. The results show the importance of primary prevention interventions, prioritizing the delay of cannabis and alcohol experimentation. Our findings can enrich the debate on drugs policies and legislation, reinforcing the need for stronger restrictions of adolescent drug access and ensuring its enforcement.