In agriculture, where the decision environment is characterized by risks and uncertainty largely due to uncertain yields and relavely low price elas cies of demand of the most commodies, decision makers require some informaon about the future and the likelihood of the possible future outcomes. Forecasng agricultural commodity price could help decision makers to take suitable decisions. This study was undertaken to forecast prices of Mustard seeds. The two types of methodological approaches such as Univariate and Mul variate techniques were adopted for analysis. In case of Univariate approach ARIMA and GARCH methods were applied whereas for mulvariate Vector Autoregressive (VAR) model was used. For forecasng Mustard prices, ARIMA (0, 1, 1) model is used which gives reasonable and acceptable forecasts; but it does not perform very well when there is volality in the data series. In this study, GARCH (1, 1) has also been a empted to forecast price but not found suitable because there was no persistent volality in the data series. In VAR model, the lag quanty arrival and lag prices in the market influence the forecasts of Mustard prices to some extent. From absolute accuracy in forecast performance measures, VAR was the most suitable forecast model for Mustard crop.
