The purpose of this study was to examine 1) whether the relationship between smoking and obesity was linear or non-linear (specifically U-shaped), and 2) if this relationship was proved to U-shaped, whether it was due to the interaction of the effects of smoking and alcohol consumption on obesity. The present study was performed using cross sectional data taken from the medical records of 400 male patients. Data was analyzed using linear and curvilinear estimation, Fisher's exact test, and two-way ANCOVA. The relationship between smoking intensity and BMI was able to be explained significantly by a quadratic model, rather than by linear model. As has been shown in other studies, this relationship was parabolic (or U-shaped), though it was not particularly remarkable. This U-shaped relationship appeared to be due to the interaction of the effects of smoking intensity and alcohol consumption on BMI. On the other hand, the relationship between smoking intensity and the percentage of body fat was able to be explained significantly by a linear as well as a quadratic model. Additionally, the interaction effect between smoking intensity and alcohol consumption did not influence the percentage of body fat significantly. In conclusion, there is no doubt that smoking and obesity are both serious health hazards. Based on the results of this and other studies, it can be confirmed that heavy smoking has a positive relationship with obesity. Therefore, the health benefits of smoking cessation, as well as the correction of unhealthy habits such as alcohol consumption should be clearly emphasized.