Background: Policies and practices related to the quasi-compulsory treatment (QCT) of substance-dependent offenders are currently implemented in many countries, despite the absence of reliable knowledge about significant predictors of treatment retention. This study aimed to identify such predictors in QCT and voluntary treatment. Methods: Participants were treated in one of 65 institutions in 5 European countries. They were interviewed at intake on substance use, crimes committed, perceived pressure for treatment, self-efficacy, stage of change, employment, and health-related variables. Binary logistic regression models were computed to identify predictors of treatment retention at an 18-month follow-up. Moderator analyses were computed to investigate whether these predictors vary by treatment condition (QCT vs. voluntary). Results: A higher number of working days in the previous month was positively associated with treatment retention, while use of heroin, crack, and multiple drugs, psychiatric problems in the previous month, and lifetime depression were negatively associated with treatment retention. Higher perceived medical pressure resulted in higher treatment retention rates only for participants in QCT. Conclusion:Predictors of substance abuse treatment retention are quite similar across both QCT and voluntary treatments. Perceived medical pressure is of higher relevance than the often-believed legal pressure for treatment retention in QCT.