Abstract
<title>Abstract</title> <p>Accurate forecasting of agricultural production is crucial for ensuring food security, optimizing resource allocation, and making informed policy decisions. This study evaluates the forecasting performance of different time series models for four major crops in Bangladesh: rice, maize, potato, and wheat. Three forecasting techniques such as Simple Exponential Smoothing, Holt’s Linear Method, and ARIMA are applied to historical crop production data. The models are assessed based on statistical performance measures, including Root Mean Square Error, Mean Absolute Error, Mean Absolute Percentage Error, and Theil's U statistic. The results indicate that Holt’s Linear Method outperforms other models for rice, maize, and potato, with optimal smoothing parameters, while ARIMA(0,3,2) is the best-performing model for wheat. A ten-year forecast (2022–2031) is generated for each crop using the best selected models, providing valuable insights into future production trends. The study highlights the importance of model selection in agricultural forecasting and suggests that policymakers and stakeholders to enhance decision-making and agricultural planning in Bangladesh.</p>