<![CDATA[One measure that can be used to see the level of farmer welfare is the farmer exchange rate (NTP), which is a comparative calculation between the price index received by farmers (IJ) and the price index paid by farmers (IB), expressed as a percentage. In reality, NTP cannot explain the actual welfare situation of farmers because the ratio value has the potential to produce biased values. Another alternative that can be used to look at farmer welfare with less potential bias is to look at the difference between the sales index and the farmer purchasing index (ID). ID data forecasting can be a reference for developing and optimizing things that need to be improved in the agricultural sector. Despite the fact that a number of external factors, such as variations in the weather throughout the year, had a significant impact on the ID value, previous research used the ARIMA model to forecast without taking exogenous factors into account. Therefore, the goal of this research is to identify the optimal ARIMAX regression model for achieving accurate forecasting results with minimal error values. This research was carried out with limitations using data from the Central Statistics Agency and the Meteorological, Climatological, and Geophysical Agency in Central Java from 2008 to 2023. The first method in this research is to prepare the data, which involved collecting secondary data such as IJ and IB along with climate data such as rainfall, duration of sunlight, air pressure, wind speed, and rice prices. Next, calculate the difference between IJ and IB to determine the ID value. Then, verify the ID data's stationarity and perform AR and MA calculations. After determining the AR and MA values, construct an ARIMAX model that incorporates external factors, search for the optimal model, and utilize the optimal model to make future predictions. The results show that the accuracy of the ARIMAX model (1,1,0) has a better value than the accuracy of the ARIMA model (1,1,0), and the results obtained in this study are better than previous studies. The authors hope that the findings of this research will serve as a benchmark for the forecasting analysis of time series data in the agricultural sector, providing the local government with a foundation for policy decisions.]]>
