This paper has presented the use of Auto Regressive Integrated Moving Average (ARIMA) method for forecasting of seasonal time series data. The dataset that has been used for modeling and forecasting is a small-scale agricultural load. ARIMA method can be applied only when the time series data is stationary. As seasonal variations make a time series non-stationary, this paper also presents analyses on testing stationarity and transforming non-stationarity into stationarity. Lastly, model has been developed with a specific selection of orders for autoregressive terms, moving average terms, differencing and seasonality and the forecasting performance has been tested and compared with the actual value. The results are encouraging, however there is scope of further research in refining the idea.