This cross-sectional study was conducted to estimate the prevalence of type 2 diabetes along with its risk factors in urban slum population of Dhaka, Bangladesh. A random sample of 1555 slum dwellers of Dhaka city (age > or = 20 years) were included in the study. Capillary blood glucose levels, fasting and 2-h after 75g oral glucose load (for a selected subjects, n = 476), were measured. Height, weight, waist and hip circumferences, blood pressure and some other important socio-demographic information on age, sex, education, income, and occupation status were collected. The overall prevalence of type 2 diabetes was found to be 8.1 percent, and the prevalence for men and women were 7.7 percent and 8.5 percent respectively. Prevalence of diabetes was found to be lower following 2-h glucose values in the selected population compared to the FBG procedure. Age, sex, literacy and waist to hip ratio for men were found as significant risk factors following both fasting blood glucose and 2-h post glucose values adjusted for a number of confounding variables. Poor to moderate agreement was observed between fasting blood glucose and 2-h glucose (kappa 0.41, p < 0.001). The agreement was even poorer between impaired fasting glucose and impaired glucose tolerance. Poor agreement between FBG and 2-h BG may raise concern for the dependability of diagnostic procedures. Higher prevalence of type 2 diabetes in the urban slum may indicate an epidemiological transition due to fast urban migration and possibly urbanization. However, this issue needs further exploration.