Epidemiologic research suggests a modest association between youth cannabis use and mood and anxiety disorders (MADs). However, current evidence is based mostly on cohort studies using data from the 20th century when cannabis was significantly less potent than today. We linked population-based survey data from 2009 to 2012 with administrative records of health services covered under universal healthcare up to 2017. The cohort included youth aged 12 to 24 years at baseline living in Ontario, Canada with no prior MAD health service use (n = 8,252). We conducted a multivariable Cox model to estimate the association between cannabis use frequency (never, <weekly, weekly+) and MAD health service use (hospitalization, ED visit, or outpatient physician visit) in the following 3 years, adjusting for sociodemographic factors and other substance use. Compared to no cannabis use, <weekly use was significantly associated with subsequent MAD health service use (aHR=1.48; 95 % CI: 1.10-1.99) while weekly+ use was not (aHR=0.92; 95 % CI: 0.63-1.34). Sex and age interactions were not significant on the multiplicative or additive scales. This study found a modest longitudinal association between youth cannabis use and MAD health service use but no evidence of a linear dose-response relationship. Further research is needed to clarify these findings.