OBJECTIVE: Despite the prevalence that smoking has declined in many countries, there is a large increase in the number of young adults starting to smoke and in per capita cigarette consumption. In some studies smoking was associated with a lower body mass index (BMI) and increased waist hip ratio (WHR). Our aim is to study the association of smoking with BMI and WHR among male adults aged 20 years and above in a community based survey as a part of the National Health Survey, 2000. METHODS: A cross sectional survey representing all parts of Oman was designed in the year 2000. A part of the survey was door to door interviews including demographic data and inquiry regarding current and former smoking for male adults aged 20 years and above. In addition, taking the weight, hip and waist measurements, blood pressure and fasting blood glucose for them. RESULTS: The crude prevalence of current smoking was 13.3% among adult males and 4.6% of them were former smokers. The mean BMI was non significantly lower among smokers than never or former smokers. There was no significant difference also regarding WHR. Adjusting BMI by 10 different multiple linear regression models for other co-variates; age, educational level, marital status, having hypertension and total fasting glucose intolerance revealed significant association in 3 of them of BMI with smoking status. Non-significant association was revealed for WHR. CONCLUSION: Current smokers were of low BMI compared to non smokers and ex smokers, and currently light smokers were also of low BMI compared to ex smokers. There was no association of central obesity to smoking status. The association between smoking status and relative weight is modified by social factors as education.