The objective of this chapter is to build the ARIMA model and forecast the in-sample and out-sample daily solar radiation data in Peninsular Malaysia. Moreover, the study also investigates the stationarity, reliability, accuracy, and performance of the model. This study involves 12 states, but Perlis�s data are removed because Perlis�s data have the same value as Kedah. The study utilizes the Box and Jenkins methodology to develop the best model for each state. Based on the three stages of the Box and Jenkins methodology, each state can be represented by the best ARIMA model. All the ARIMA models also produced smaller error values which indicate the fitted or forecasted values follow the same trend as the actual data. © 2021, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.