Regional climate model of the third generation (RegCM3) has been experimentally used in seasonal simulation of Rainfall in Pakistan using ERA-40 reanalysis of ECMWF as lateral boundary and initial conditions for driving the RegCM3. The performance of Regional Climate Model (RegCM3), version released in February` 2005, for Pakistan is examined through experiments to simulate the amount of total precipitation in monsoon of the year 1992, with model resolution of 60km and domain of about 50°E- 100°E, 05°N - 45°N. The model output is then compared with real time and Climatic Research Unit (CRU) observational grid Data. In this study, experiments have been done using Grell Convective Closure Scheme with 1-Arakawa & Schubert (AS), 2-Fritsch & Chappell (FC) closure assumptions. The effect of these convective closure assumptions has also been examined. It is found, in this study, that the simulation of summer monsoon rainfall over Pakistan is reasonable. The amount of simulated rainfall with Fritsch & Chappell closure is more than that with Arakawa & Schubert closure assumption in this experiment, over Pakistan. This study also revealed that the average amount of simulated rainfall, with both closure assumptions, is 51% approx. less than that calculated from Climatic Research Unit (CRU) data and almost 71% approx. less than the real time data.