Background <p>Many countries and jurisdictions have recently adopted more lenient cannabis laws. While use remains prohibited in Sweden, this stance is debated due to high drug-related mortality. To inform this debate, we predicted changes in cannabis use following a hypothetical decriminalization in Sweden.</p> Methods <p>Jurisdiction-level data on self-reported cannabis use from 12 countries (across Europe and Australia) and four U.S. states were used in a multilevel meta-regression model to predict the most likely changes in past 12-month and past 30-day cannabis use following decriminalization in Sweden.</p> Results <p>We predicted an immediate increase in the prevalence of both past 12-month and past 30-day cannabis use following decriminalization. However, the longer-term trends differed between these measures. For past 12-month use, the gap in prevalence between scenarios with and without decriminalization gradually narrowed over time. In contrast, for past 30-day use, the gap widened.</p> Conclusions <p>Decriminalizing cannabis use in Sweden would lead to an initial increase of self-reported cannabis use. The extent to which this reflects a genuine increase in use or an increased propensity to disclose the behavior remains to be understood. After this initial increase experimental use tends to stabilize, while recurrent use may continue to increase, probably indicating an increasing pool of individuals transitioning towards dependence. The method we propose for predicting population trends in cannabis use following decriminalization can be easily replicated in other context and used to support evidence-based policy decisions.</p>