A study was conducted on modelling and forecasting the time series data of total herbicide consumption in India. Among many time series methodology, Box-Jenkins Autoregressive Integrated Moving Average (ARIMA) model was used for modelling and forecasting purposes using data from 1990 to 2010. Before the modelling, stationarity of the data was checked using Augmented Dicky Fuller test. Best model was chosen using two criterion viz. Akaike information criterion and Schwarz’s Bayesian criterion. ARIMA (0, 1, 1) model was found to the best among many models from ARIMA family. Forecasting was done using the best model and prediction for total herbicide consumption in India was made for next three year (2011, 2012, 2013) as 6624, 6581 and 6562 tonnes respectively.