BACKGROUND : Long-term benzodiazepine (BZD) use among seniors is mostly inappropriate and associated with adverse health outcomes. To prevent these consequences, withdrawal is crucial, yet knowledge is limited about what predicts BZD discontinuation. Until now, most studies have focused on sociodemographic and BZD intake factors as predictors while neglecting psychological factors. This research addresses this issue by studying how the intensity of depressive symptoms, social support satisfaction, self-perceived competence in the ability to withdraw, and overall quality of sleep predict discontinuation in long-term older consumers. METHODS : Seventy-three participants aged 60 years and older were enrolled in this study. There were four time measures: before discontinuation (T1), after (T2), 3 months after (T3), and 12 months after (T4). Data were collected in the "Programme d'Aide du Succès au SEvrage" (PASSE-60+) study. RESULTS : Social support satisfaction predicted discontinuation at T2 and T4. Self-perceived competence in the ability to withdraw and depressive symptoms predicted discontinuation at T4. This later prediction was counterintuitive; higher depressive symptoms at T1 were linked with higher discontinuation success. BZD intake factors (length of use and dose) were good predictors for short term discontinuation. Psychological factors were moderate predictors for short term and good predictors for long term discontinuation. CONCLUSIONS : Psychological factors are good predictors of discontinuation and are better predictors than BZD intake factors of long-term discontinuation. Discontinuation programs should focus on social support and self-perceived competence to improve their efficacy. Further studies are needed to acquire a more complete picture of the psychological predictors of discontinuation success. ClinicalTrials.gov Identifier: NCT02281175.