Paranoia is characterized by a lack of perceived social safeness and associated negative affect. Low self‐esteem, negative self‐concepts and negative emotions have been shown to contribute to paranoid symptom formation. Thus, interventions focusing on affiliation and positive affect might be particularly helpful for patients with paranoia. The present study experimentally tested the effect of a one‐session, brief compassion‐focused imagery derived from Compassion‐Focused Therapy (Gilbert, ) versus a control imagery condition in a repeated measures randomized design. A negative affective state was induced via in‐sensu exposure to a recent distressful social situation in order to provide a minimum level of threat‐related arousal to be down‐regulated by the interventions thereafter. The sample consisted of psychotic patients with paranoid ideation (N = 51) who were randomly assigned to one of the experimental conditions. Effects on postulated causal mechanisms, i.e., self‐relating (self‐reassurance, self‐compassion, self‐criticism), and affect (self‐reported affective states, skin conductance levels) as well as on state paranoia, were tested. Subjective benefit and appraisals of the intervention were explored. There were no specific intervention effects on negative self‐relating, negative affect and skin‐conductance or on paranoia. However, compassion‐focused imagery had significant effects on self‐reassurance and happiness. Explorative analyses revealed that the majority of the participants appraised the intervention in a positive manner, indicating good acceptance. The intervention showed an effect on some of the postulated mechanisms but not on others, which might have been because of its brevity. Further investigation of interventions targeting affiliation for people with paranoid experiences appears worthwhile. Copyright © 2016 John Wiley & Sons, Ltd. Key Practitioner Message Affiliative imagery work is feasible and appraised positively in psychotic patients. Brief compassion focused imagery increased feelings of happiness and reassurance but did not improve negative self‐relating, negative affect or paranoia. Further investigation is warranted to identify which patients benefit most from affiliative imagery.