Objective: The study was designed to estimate the prevalence oftype 2 diabetes among the urban slum population and to makea valid comparison of differential prevalences along with itsrisk factors. Further, the agreement between of FBG and 2-hBG were also examined.Methods: The study utilized two sets of data, one including the prevalence of type 2 diabetes among the urban slum population in Dhaka city, and a previous study conducted in selected rural areas. The rural study was performed among 5000 individuals (aged >20years) both males and females in 1999. The urban study was conducted among the urban slum dwellers (migrants) from those specific rural areas. A cross-sectional study was conducted among 1555 slum population both male and females aged >20 years in 2001.Capillary fasting blood glucose levels, and 2-h post glucose test after a 75-g glucose drink were measured for a number of selected subjects from both urban (n=476) and rural(n=1046) population. Important anthropometrical indicators (Height, weight, waist and hip circumference) including blood pressure and socio-demographic information werecollected.Results: A higher prevalence of diabetes was found among the urban subjects 8.1 percent compared to rural population 2.3 percent. The study population was lean both for urban and rural with mean BMI (rural 20.2 and urban 19.4). Female had higher prevalence of diabetes compared to male both in urban and rural (urban female-8.5%, male 7.7% and rural female-2.5%, male-1.9%). Age, sex and waist to hip ratio for male were found to be significant risk factors following FBG and 2-h glucose values adjusted for a number of confounding variables. Poor agreement was observed of between FBG and 2-h BG values.Conclusion: Increased risks for the development of diabetes were observed among the urban subjects compared to its source population. The risk factors were mostly similar for both urban and rural subjects. This may indicate that we are representing a unique form of type 2 diabetes in our lean population without obesity and hypertension. Applicability of universal cut-off points for obesity status may call for an examination in order to classify the people at risk. Further FBG did not appear to provide an under estimation ofDM prevalence compared to2-h BG.