Background. Cognitive dysfunction is a common feature in alcohol use disorders. Its persistence following alcohol detoxification may impair quality of life and increase the risk of relapse. We analyzed cognitive impairment changes using the Montreal Cognitive Assessment (MoCA) score in a large sample of alcohol-dependent inpatients hospitalized for at least 4 weeks. Method. This was an observational longitudinal survey. Inclusion criteria were alcohol dependence (DSM-IV) and alcohol abstinence for at least one week. The MoCA test was administered on admission and at discharge. Results. 236 patients were included. The mean MoCA score significantly increased from 22.1 ± 3.7 on admission to 25.11 ± 3.12 at discharge. The corresponding effect-size of improvement was high, 1.1 [95% CI 1.0-1.2]. The degree of improvement was inversely correlated with the baseline MoCA score. The rate of high and normal, that is, >26, MoCA values increased from 15.8% on admission to 53.8% at discharge. MoCA score improvement was not correlated with the total length of abstinence prior to admission. Conclusion. The MoCA score seems to be a useful tool for measuring changes in cognitive performance in alcohol-dependent patients. A significant improvement in cognitive function was observed whatever the degree of impairment on admission and even after a long abstinence period.