<p>In recent years, an increasing number of graph layout algorithms have emerged. A prominent approach for comparing these algorithms and assessing layout quality is to employ aesthetic metrics. However, traditional methods for computing these metrics typically rely on predefined formulas, which are computationally slow. In this paper, we introduce a graph neural network-based model that utilizes vertex coordinates and edge attributes to derive representations containing both local geometric details and global layout topology. These representations are then processed through a series of multilayer graph convolutions, pooling operations, and multilayer perceptrons to predict graph layout quality. Our model not only achieves high evaluation performance but also demonstrates remarkable efficiency, rendering it suitable for layout evaluation tasks on large graphs. To validate its efficacy, we conduct extensive quantitative evaluations and compare it against existing models. The experimental results affirm the efficiency and accuracy of our proposed model .</p> Graphical abstract <p></p>