Considering the effective service life characteristics of products in the background of product sales, this study first designs a generalized fractional-order accumulation generation matrix covering the effective accumulation percentage, replaces this matrix with the first-order accumulation generation operation in classical GMP(1,1), and establishes the definition equation of fractional-order accumulation GMP(1,1) [abbreviated as GFAGMP(1,1,D)] model. Second, the study derives the level ratio of the model via two other forms of the model [GFAGMP(1,1,M) and GFAGMP(1,1,H)] using integral and power function transformation, which is combined with the power function transformation and the relationship between the GFAGMP(1,1) model and the generalized the GM(1,1) model. The optimization model of accumulation generation matrix order $r$, effective accumulation percentage index $p$. and power exponent $m$ . is established, and the parameter optimization of the model is completed through particle swarm optimization. Finally, the simulation results of different grey models are compared with the actual sales data of Chinese refrigerators. This comparison demonstrates the feasibility and effectiveness of the GFAGMP(1,1) model in forecasting the home-appliance supply chain demand in China.