Abstract: A key property of grey prediction models is their unbiased nature, which focuses on eliminating discontinuity between differencing and differentiation during their construction. Meanwhile, an unbiased proof is required to ensure this property. In this paper, an unbiased EDGM(1,1) model incorporating Euler polynomials is constructed. Firstly, this model integrates Euler polynomials and a time disturbance parameter into the classical GM(1,1) model, enabling this EDGM(1,1) model to flexibly handle sequences with various data characteristics. Subsequently, the difference and its discrete forms are derived, and the latter is then solved using the least squares method and mathematical induction. Then the Particle Swarm Optimization (PSO) algorithm is implemented to improve the model's parameters. Secondly, the compatibility of the EDGM(1,1) model is demonstrated, and its unbiasedness towards three characteristic sequences is proven based on Cramer's rule. Finally, the performance of the EDGM(1,1) model is evaluated comprehensively against five competing models using three metrics: MAPE, RMSE, and R². The comparative analysis shows that the EDGM(1,1) model outshines other models in robustness and accuracy. Furthermore, the novel model is designed to forecast China's primary energy outputs, aiming to provide references for energy policies and decision-making. 

Key words: Grey forecasting model, Euler polynomial, Unbiasedness, Primary energy production forecasting