With the process of urbanization, the prediction of traffic information assumes considerable significance for the effective management of traffic. Considering the temporal characteristics is a universal practice when making predictions about traffic information. This method uses the Gated Recurrent Unit (GRU) to get temporal features related to traffic flow, thereby revealing the traffic flow of the road. Nevertheless, this approach fails to incorporate the spatial structural attributes of the road network. While some scholars have considered the regional structural features of road networks, their work has typically focused on extracting Euclidean structural features and has paid little attention to non-Euclidean structural features. This article proposes a method for predicting traffic flow information. It uses GRU to obtain temporal characteristics about traffic flow and an attention mechanism to assign different weights. Subsequently, a graph convolutional network (GCN) is employed for the extraction of non-Euclidean structural features of the road network space. Afterwards, the efficacy of this method is demonstrated through an evaluation against a series of baseline models. Experimental results show that the proposed method produces more accurate predictions than the baseline model on both PEMS_BAY and METR_LA datasets.