<![CDATA[General Background Sugar represents an essential commodity supporting public food security and industrial production in Indonesia. Specific Background PT. PG Candi Baru faces annual sales fluctuations of SHS 1 sugar, where inaccurate sales estimation risks increasing production expenditures and diminishing investment efficiency. Knowledge Gap Comparative sales prediction models utilizing empirical historical data remain underanalyzed for operational planning in regional sugar manufacturing plants. Aims This study evaluates and compares the accuracy of the Moving Average and Naive methods to determine the most precise forecasting model for sugar sales. Results Quantitative evaluation via POM QM for Windows demonstrates that the 6-year Moving Average yields lower error rates, with Mean Absolute Deviation (MAD) of 1.726, Mean Squared Error (MSE) of 5.661, and Mean Absolute Percentage Error (MAPE) of 8.45%, projecting subsequent annual sales at 21,781 tons. Novelty The investigation provides an empirical performance baseline directly contrasting time-series smoothing against naive baseline projections within a state-affiliated sugar refinery setting. Implications Adopting the Moving Average approach enables management to execute precise operational planning and mitigate financial risks associated with supply-demand imbalances. Key Findings Highlights The 6-year Moving Average method delivered higher accuracy than the Naive method, recording lower error metrics across MAD, MSE, and MAPE. The projected annual sugar product sales for the upcoming period at PT. PG Candi Baru reaches 21,781 tons. Incorporating time-series forecasting metrics provides strategic support for production planning and operational cost control. Keywords Forecasting, Moving Average, Naive Method, Sales Prediction, Sugar Industry]]>
