Background: The clinical literature cautions against use of meditation by people with psychosis. There is, however, evidence for acceptance-based therapy reducing relapse, and some evidence for clinical benefits of mindfulness groups for people with distressing psychosis, though no data on whether participants became more mindful. Aims: To assess feasibility of randomized evaluation of group mindfulness therapy for psychosis, to replicate clinical gains observed in one small uncontrolled study, and to assess for changes in mindfulness. Method: Twenty-two participants with current distressing psychotic experiences were allocated at random between group-based mindfulness training and a waiting list for this therapy. Mindfulness training comprised twice-weekly sessions for 5 weeks, plus home practice (meditation CDs were supplied), followed by 5 weeks of home practice. Results: There were no significant differences between intervention and waiting-list participants. Secondary analyses combining both groups and comparing scores before and after mindfulness training revealed significant improvement in clinical functioning (p = .013) and mindfulness of distressing thoughts and images (p = .037). Conclusions: Findings on feasibility are encouraging and secondary analyses replicated earlier clinical benefits and showed improved mindfulness of thoughts and images, but not voices.