Based on the legal system of enterprises, in the realm of intelligent financial management, accurately predicting the impact of retail promotional activities on revenue is crucial. However, traditional methods often struggle to address the complexity of promotional effects and the variability of consumer behavior, resulting in less accurate predictions. To tackle this issue, we propose a novel model, TCGNet, based on deep learning techniques. This model integrates TCN and GNN, optimized by the GWO, to provide more precise predictions of the impact of promotional activities on revenue. Experimental results demonstrate significant advantages of the TCGNet model across various metrics, offering more reliable support for retail financial management. This research aims to furnish businesses with more precise and timely financial planning and budget management support, with the potential to aid in the formulation of more effective marketing strategies and financial decisions.