This paper predicts the urban male population of Bangladesh using geometric growth rate method. The predictions are computed in three stages. In the first stage, the urban male population for the years 1981 and 1991 was predicted using the smoothed data for those years. The first stage predictions were obtained by using generalized negative exponential model estimated by nonlinear least squares method. While the urban male population for the year 2001 was predicted by a linear model. Using the cross validation predictive power (CVPP) and 2 R this article constructed shrinkage coefficient which determines the adequacy of the first stage predicted values. These predicted values are then used in the second stage to estimate the geometric growth rate for different age groups. Finally, considering year 2001 urban male population as the base period and using the estimated geometric growth rate of the second stage the predictions of the urban male populations are computed for 2002 through to 2031.