Objective: Despite a large evidence base indicating that cognitive skills programs can reduce reoffending in individuals without mental illness, there have been no randomized controlled trials (RCTs) to determine their effectiveness in mentally disordered offenders (MDOs). In the first RCT of a cognitive skills program for MDOs, we aimed to investigate whether participation in the Reasoning and Rehabilitation (R&R) program reduced violence and antisocial behavior in this population. Method: Eighty-four male inpatients with a psychotic disorder and a history of violence were recruited from medium secure forensic hospitals. Participants were randomized to receive the R&R program, consisting of 36 two-hour sessions, or treatment as usual (TAU). Incidents of violence and antisocial behavior (verbal aggression, substance use, and leave violations) were assessed during treatment and at 12-months posttreatment. Results: Relative to the TAU group, incident rates of verbal aggression and leave violations during the treatment period were significantly lower in the R&R group; the effect on verbal aggression was maintained at 12-months posttreatment. Half of those randomized to receive R&R did not complete treatment; post hoc analyses were therefore conducted to compare treatment responses in program completers and noncompleters. After controlling for psychopathic traits, incidents of violence, verbal aggression, and leave violations during treatment were significantly lower in program completers, and there were significant effects of program completion on verbal aggression and substance use at 12-months posttreatment. Conclusions: R&R leads to a reduction in incidents of antisocial behavior in MDO populations, with potentially greater impact on those who complete treatment. (Contains 3 tables and 1 figure.)