Forecasting of prices of commodities specially those of agricultural commodities is very difficult because they are not only governed by demand and supply but by so many other factors which are beyond control like weather vagaries, storage capacity, transportation etc. In this paper times series namely ARIMA (Autoregressive Integrated Moving Average) methodology given by Box and Jenkins has been used for forecasting prices of edible oils and this approach has been compared with ANN (Artificial Neural Network) methodology.