Many analysis, correlation and forecast related to sea surface temperature (SST) or air temperature have been surveyed by researchers in the past. Temperature can affect many aspects in our daily life. In Hokkaido, where fishing industry is the largest among Japan’s fishing industry, SST can affect quantity and different types of species of fishes. Therefore, accurately knowing and anticipating the temperature can help fishermen to strategize their fishing plans better.

In this paper, a forecast of monthly average sea surface temperature using seasonal ARIMA (Autoregressive Integration Moving Average) models in time series is performed. Many ARIMA models are derived and compared before arriving at a final appropriate and adequate model. After thorough analysis (i.e. identification, evaluation and validation) of various models, an appropriate and the best seasonal ARIMA model, which is ARIMA(2,0,1)(1,1,2)[12], is selected and recommended. Subsequently, the chosen seasonal ARIMA model is used to implement the forecast and the result of the forecast shows that the prediction is very accurate and the seasonal ARIMA(2,0,1)(1,1,2)[12] model is a very good model.