Aims. To estimate the life-time prevalance rate of tobacco/nicotine dependence and demographic variables and smoking habits associated with the disorder in male ever-smokers in Japan. Design. A cross-sectional community-based interview study. Setting. Takayama city, Gifu prefecture, Japan. Participants. A total of 170 male ever-smokers aged 35 years or older selected randomly from a community in Japan were interviewed. The response rate was 85%. Measurements. The WHO Composite International Diagnostic Interview (CIDI) was used to make diagnosis of tobacco/nicotine dependence according to ICD-10, DSM-III-R and DSM-IV. The Fagerstrom Tolerance questionnaire (FTQ) was also administered and those who had a FTQ score of 7 or above were identified. Findings. The life-time prevalence rates of tobacco/nicotine dependence in male ever-smokers were 42%, 26% and 32% according to ICD-10, DSM-III-R and DSM-IV criteria, respectively; 19% had a FTQ score of 7 or above. The ICD-10 diagnosis was significantly and negatively associated with quitting smoking (p < 0.05). Multiple logistic regression analyses indicated that number of cigarettes per day when they smoked the most was significantly associated with higher life-time risks of the disorder according to DSM-III-R, DSM-IV and Fagerstrom's classification (p < 0.05). The length of cigarette smoked was associated with higher life-time risks of ICD-10 and DSM-IV diagnoses, and years of smoking were associated with higher life-time risks of ICD-10, DSM-III-R and DSM-IV diagnoses (p < 0. 05). Younger birth cohorts had higher cumulative rates of the disorder according to DSM-IV (p for trend < 0. 05). Conclusions. Life-time prevalence rates of tobacco/nicotine dependence according to ICD-10, DSM-III-R and DSM-IV in male ever-smokers in Japan were within the range of rates reported in previous US studies; rates of FTQ score of 7 or above were lower. Fagerstrom scores and diagnostic criteria appear to reflect different aspects of dependence.