Insomnia in psychiatric patients is frequently underestimated in clinical practice. Usually drugs are prescribed for the treatment of this disorder but non-pharmacological intervention can be successfully used. The present study aimed at evaluating the efficacy of a two-session psychoeducational intervention in improving persistent non-organic insomnia and reducing the administration of PRN therapy in severely mentally ill patients. A pre-post study was performed on 36 psychiatric patients admitted to a residential psychiatric unit. The Nocturnal Sleep Onset Scale (NSOS) and Daytime Sleepiness Scale (DSS), the sleep onset latency, the time awake after sleep onset and the numbers of awakenings were gathered 2 weeks before the intervention (T0), immediately prior the intervention (T1), 2 weeks after the last session of the intervention (T2) and a 3-month follow-up (T3). The total number of administrations of PRN therapy from T0 to T1 and from T1 to T2 were also examined. A significant reduction was shown on the NSOS, the sleep onset latency and in the time awake after sleep onset from T1 to T2 and from T1 to T3, while no significant difference was found between T0 and T1. A significant decrease on the mean number of administrations of PRN therapy was also found between 15 days before the intervention (T0-T1) and 15 days after intervention (T1-T2). The initial results of this study seems to suggest the possible efficacy of a short-term psychoeducational intervention on improving persistent non-organic insomnia in severely mentally ill patients. Further control studies are necessary to confirm these findings.