The study on demand forecasting is pertinent in countries like Bangladesh where the industry specifically heavy industries are dying for accurate demand forecast. Autoregressive Integrated Moving Average (ARIMA) model has been used to detect patterns, trends and seasonal influences in order to generate accurate short-term forecasts. Data has been collected monthly, ranging from July to June of each corresponding year. Data processing involved zero-value occurrences, time-series decomposition and performed tests of stationarity. SARIMA with Exogenous Model, ARIMA Model for VIX Index on overridden series have been developed and, forecast performance was evaluated. Moreover, prediction range by specifying confidence interval levels has been prepared. It is found a strong monthly seasonal effect of production, with peaks in the first quarter and minimum values in July-August compatible with ideal rain conditions. The start of a slight upward trend in overall production indicated slow sector expansion. The SARIMA model selected had an average forecast accuracy of MAPE=34.84% and RMSE=16,349 MT. The ambiguity reinforces the nuanced set of variables which determine if heavy industries like Steel re-rolling mills in Bangladesh are allowed to produce.