Quality assessment of tobacco leaf aging is a fundamental research priority in the tobacco industry. To address industry requirements, this study systematically selected representative samples from production data and constructed a specialized aging dataset through rigorous pre-processing. Based on this dataset, we developed an innovative classification model for tobacco leaf aging stages and implemented a sophisticated image feature fusion approach. By integrating multibranch feature information, the model effectively captures and synthesizes key aging characteristics, significantly enriching feature representation and improving its ability to distinguish between aging stages. Experimental results show that our model achieves an accuracy of 88.67% in aging stage classification, significantly outperforming baseline models. These findings establish a robust technical framework for precisely identifying tobacco leaf aging stages.