BACKGROUND: Negative beliefs about the self, including low self-compassion, have been identified as a putative causal factor in the occurrence of paranoia. Therefore, improving self-compassion may be one route to reduce paranoia. AIMS: To assess the feasibility, acceptability, and potential clinical effects of a brief compassionate imagery intervention for patients with persecutory delusions. METHOD: Twelve patients with persecutory delusions received an individual four-session compassionate imagery intervention. Assessments of self-concept and paranoia were completed before treatment, immediately after treatment, and at 1-month follow-up. A qualitative study exploring participants' experiences of the treatment was also completed. RESULTS: Twelve out of 14 eligible patients referred to the study agreed to take part. All participants completed all therapy sessions and assessments. Post-treatment, there were improvements in self-compassion (change score -0.64, 95% CI -1.04, -0.24, d = -1.78), negative beliefs about the self (change score 2.42, 95% CI -0.37, 5.20, d = 0.51), and paranoia (change score 10.08, 95% CI 3.47, 16.69, d = 0.61). There were no serious adverse events. Three themes emerged from the qualitative analysis: 'effortful learning', 'seeing change' and 'taking it forward'. Participants described a process of active and effortful engagement in therapy which was rewarded with positive changes, including feeling calmer, gaining clarity, and developing acceptance. CONCLUSION: This uncontrolled feasibility study indicates that a brief compassionate imagery intervention for patients with persecutory delusions is feasible, acceptable, and may lead to clinical benefits.