Among university students, large numbers are involved with smoking and suffer from many chronic diseases. This study examined tobacco smoking prevalence and potential predictors among university students in Sylhet Division, Bangladesh. A total of 416 students were selected for face-to-face interviews. Logistic regression analysis was used to fulfil the specific objectives of the study. It was discovered that the prevalence of smoking among university students in Sylhet Division was 37% and almost half of current male students were smokers. Mother's occupation and peer smoking status were found to be significant factors. Although data were collected using multistage sampling, the stages (type of university, universities, departments and academic levels) were insignificant with smoking status. There is a 2.1 times greater likelihood of a student becoming a smoker if a close friend is a smoker. The adjusted living expenses (>$100) was a proactive factor, though individually it was a significant factor. This study and previous studies found that smoking initiation among female students is negligible, so gender-specific health promotion and intervention is needed in Bangladesh. Tobacco control awareness programs and 'No smoking' signage should be displayed within university campuses and sanctioned by university authorities.