The diagnosis of personality disorders in learning disability is often contentious. Some studies showing exceptionally high prevalence rates for these disorders have raised questions about its clinical usefulness in learning disability. This study aims to estimate the prevalence of a clinical diagnosis of personality disorder among those with learning disability and mental health or behavioural problems and describe their characteristics. People with a clinical diagnosis of personality disorder were identified by screening the psychiatric casenotes of patients from two areas of Leicestershire. ICD-10 DCR (WHO, 1993) criteria were then used to identify the type of personality disorder. Socio-demographic and clinical details were collected using a semi-structured proforma. Twenty nine out of the 430 patients seen by two consultant psychiatrists had a clinical diagnosis of personality disorder, thus giving a prevalence of 7%. The vast majority had either mild or borderline learning disability. Only 34% had a co-morbid mental illness, but the vast majority were on psychotropic medication. Fifty nine percent had an history of hospital admissions and the same proportion needed active support from an outreach team. Some recent studies in learning disability have suggested prevalence of up to 90% for personality disorders. The prevalence figure of 7% from this survey suggests that in routine clinical practice, the diagnosis is limited to those with mild learning disability and significant behavioural disruption. Those diagnosed in this manner also satisfy the rigorous ICD 10-DCR criteria (WHO, 1993). To diagnose personality disorders across the whole range of learning disability, diagnostic criteria that are specific for different developmental levels are needed. In the absence of such norms, the current clinical practice of limiting the diagnosis of personality disorders to a tightly defined group may make more sense, as it identifies a group of people with specific care needs and resource implications.