Background:Auditory hallucinations may persist in a subset of chronic psychotic patients in spite of treatment. It is important to understand the personal meaning and significance of voices in these patients. In spite of its relevance, only a limited literature is available.Aim:This exploratory study aimed to assess the beliefs regarding voices in treatment-seeking patients with chronic schizophrenia having persistent auditory verbal hallucinations (AVHs) and assess their relation to the severity of psychosis.Materials and Methods:We recruited thirty adult patients with chronic schizophrenia as per the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition with both lifetime and current significant AVHs (≥50% days/month). Co-occurring psychiatric disorders were ruled out using the Mini International Neuropsychiatric Interview-7.0.0. Patients were assessed using a semi-structured proforma, Beliefs about Voices Questionnaire-Revised (BAVQ-R), Psychotic Symptom Rating Scale (PSYRATS), Scale for the Assessment of Positive Symptoms (SAPS), Scale for the Assessment of Negative Symptoms (SANS), and Clinical Global Impression-Schizophrenia (CGI-SCH)-severity.Results:The median age of the patients was 32 years (interquartile range [IQR]: 23.8–40.5). The median duration of illness and treatment was 7 years (IQR: 3.4–15.0) and 3 years (IQR: 1.9–10.5), respectively. Higher BAVQ-R scores were found on “malevolence,” “omnipotence,” and “emotional and behavioral resistance.” These beliefs had a significant positive correlation with PSYRATS hallucination subscale, but not with the severity of psychosis (SAPS, SANS, and CGI-SCH). The sample had lower scores for “benevolence” and “engagement” subscales of BAV-Q.Conclusion:Overall, the study sample believed AVH to be more malicious and omnipotent rather than benevolent, and resisted the voices, engaging only minimally with them. These beliefs were not related to the severity of psychosis, but were related to the severity of hallucinations. Assessing the beliefs regarding AVH in larger, diverse samples may help to plan behavioral interventions.