Introduction Social disparities in health is becoming a topic of major importance. Recent studies presents data on social differences in morbidity on various diseases. This paper shows how mental health disorders prevalences are strongly linked to sociological variables. It is based on the data of the research-action entitled « Mental Health in General Population: images and realities (MHGP) » carried out by the World Health Organisation Collaborative Centre (Lille, France) and the Direction of research, studies, assessment and statistics (Drees) of the French Ministry of Health, in a sample of 36 000 French subjects over 18 years old, between 1999 and 2003. Objectives The MHGP Survey is an international multisite study aimed at assessing the prevalence of major psychiatric disorders in the general population, as well as describing the representations attached to insanity, mental illness and depression, and the related care. It allows the analysis of the impact of professional situation on prevalence data in a large sample. Methods The MHGP Survey was carried out in 47 French public sites between 1999 and 2003. A face-to-face questionnaire was used to interview a representative sample of French metropolitan subjects, aged 18 and over, non-institutionalized and homeless. These subjects were recruited using quota sampling for age, gender, socio-professional and education levels, according to data from the 1999 national French population census. Representations of insane, mentally ill and depressive persons were explored by a specific questionnaire with open and semi-open questions. Psychiatric diagnoses were identified using the Mini International Neuropsychiatric Interview (MINI); including: mood disorders, anxiety disorders, alcohol and drug disorders, psychotic disorders and suicidal risk. A national database was then constituted by pooling data from all sites, weighted for age, gender, level of education, socio-professional level and work status to be representative of the French general population. Prevalence rates of mental disorders are very similar to those observed in other studies. Social disparities were analysed mainly using the work status categories, comparing employed people and unemployed people (unemployed, retired, students, housewife). Results Results show that mental disorders and suicidal risk are not fairly distributed between the various social categories. Half of unemployed people suffers from at least one mental disorder, whereas less than a third of employed people. For example, the prevalence of mood disorders is more than twice higher in unemployed (23,6%) than in employed (11,7%). Suicidal risk is specially high in unemployed people (22%) and students (15%). As in other pathologies and in causes of death, presence of mental disorder or suicidal risk negatively correlates with social status. Prevalence of mental disorders is higher in unemployed people and unskilled workers than in executives, managers and intellectual professions. Social disparities in mental health do not distinguish with those in health in general. Therefore, the issue of search for the causes of disparities is similar to that for any disparity in health.