Economic forecasting is a rapidly developing field with wide applicability in business and government. In the past years we see that studies on forecasting macroeconomic variables in Bangladesh rarely exist. In this study, we used ARIMA and VAR model to analyze and forecast GDP with four of its sectors where data are extracted from Bangladesh Bureau of Statistics (BBS) from the year 1979-80 to 2007-08 at constant price (base year 1995-96). This study made a comparison of these two forecasting techniques to see which suits better in forecasting the GDP and its sectors. The empirical findings of this study suggest that in forecasting the sectors of GDP of Bangladesh VAR analysis is more efficient tools over ARIMA. Especially for multivariate forecasting analysis VAR is the right technique. ARIMA model plays quite an excellent role in univariate econometric time series data analysis of GDP of Bangladesh and its sectors.