为识别无人机与有人机融合运行空域的关键节点，支撑空域的安全管理与网络优化，本文构建了一种能同时刻画空域网络结构特性与交通运行功能特性的关键节点识别与高效预测方法。首先，从网络结构-功能耦合视角出发，构建融合全局拓扑结构与载流性能的加权网络效率指标；并通过刻画节点失效引起的加权网络效率损失率，建立关键节点重要度量化模型，实现对节点结构重要性与运行影响力的统一刻画。在此基础上，针对大规模融合运行空域中关键节点重要度计算复杂度高的问题，构建了一种基于注意力机制的图卷积网络（AGCN）预测模型，对节点重要度进行快速估计，并与CNN和GCN模型进行对比验证。实验结果表明，所提出的AGCN模型能准确预测节点重要度分布，在测试集上，其平均绝对误差和均方根误差相较于性能最优的GCN模型分别降低4.04%和6.12%，计算时间也显著减少。研究表明，结构-功能耦合的重要度定义能刻画节点失效对融合运行空域整体效率的影响差异，基于该指标，节点关键性排序呈现出一致性与区分性特征；同时，AGCN模型可在降低计算复杂度的情况下有效估计节点重要度，适用于大规模融合运行空域的关键节点分析。

To identify key nodes in the integrated operational airspace involving unmanned and manned aerial vehicles, and support the safety management and network optimization of airspace, this study develops a key node identification and efficient prediction method that simultaneously characterizes the structural properties of airspace networks and the functional characteristics of traffic operations. From a structure-function coupling perspective, a weighted network efficiency index integrating global topological structure and traffic-carrying capability is constructed. By quantifying the loss rate of weighted network efficiency induced by node failures, a quantitative model for critical node importance is established, enabling a unified characterization of nodes′ structural importance and operational influence. On this basis, to address the high computational complexity of critical node importance evaluation in large-scale integrated airspace networks, an attention-based graph convolutional network (AGCN) prediction model is developed to enable rapid prediction of node importance, and its performance is comparatively evaluated with convolutional neural network (CNN) and graph convolutional network (GCN) models. Experimental results demonstrate that the proposed AGCN model can accurately predict the distribution of node importance. On the test dataset, the mean absolute error (MAE) and root mean square error (RMSE) of the AGCN model are reduced by 4.04% and 6.12%, respectively, compared with the best-performing GCN model, while computational time is also significantly reduced. The results indicate that the structure-function coupled importance definition effectively captures the differential impact of node failures on overall operational efficiency in integrated airspace, and the resulting node criticality rankings exhibit both consistency and discriminability. Moreover, the AGCN model enables effective estimation of node importance with reduced computational complexity, making it suitable for key node analysis in large-scale integrated airspace operations.