Aims: It is well established that tobacco use is associated with mental disorders. However, the association between tobacco use and mental disorders has not yet been examined sufficiently in prospective, population-based studies. The current study is aimed at examining whether smoking is associated with first-ever incidence of mental disorders. Design, setting and participants: We conducted a prospective, population-based epidemiological study (the Netherlands Mental Health Survey and Incidence Study: NEMESIS) in which a representative sample of adults aged 18-64 years (n = 7076) were interviewed to establish the presence of a broad range of mental disorders. We re-interviewed them at 1 year (n = 5618; response 79.4%) and 3 years (n = 4796; 67.8%) after baseline. Measurements: The presence of mental disorders was assessed according to DSM-III-R criteria with the Composite International Diagnostic Interview (CIDI), which was administered by trained lay interviewers. Tobacco use was assessed by asking respondents whether they had smoked in the past year, and how many cigarettes they smoked. Findings: Subjects who smoked but never had a mental disorder in their life, had an increased risk of developing a mental disorder (P < 0.01), and this remained significant after correcting for major risk indicators of mental disorders. Conclusions: Smoking is associated not only with the prevalence, but also with first-ever incidence of mental disorders. More research is needed to study the causal pathways. © 2007 The Authors.