A small controlled trial of cognitive behaviour therapy for drug-resistant psychosis is reported. The study was designed as a pilot study for a future larger and longer randomized controlled trial. The therapy was offered to patients with a diagnosis of schizophrenia or schizo-affective psychosis who presented unremitting positive symptoms. An average of 16 sessions were delivered over a six-month period. The results of this pilot study are promising. Rates of engagement in therapy were high. The treatment group also improved significantly on a number of key symptom measures when compared with the controls. These were reductions in delusional conviction, general symptomatology and depression scores. Future studies should offer therapy over a longer period, targeting social as well as symptom change, and considering factors which will enhance maintenance of improvement.