In the realm of modern service computing, delivering personalized and precise service recommendations to users has emerged as a critical challenge and garnered significant attention. Drawing inspiration from recommender systems, existing Quality of Service (QoS) prediction methods predominantly utilize collaborative filtering (CF), low-rank decomposition, and neural networks to process raw data and generate accurate predictions. However, their performance is hindered by insufficient exploration of features. To address this, we propose the HCE-RSA (High-order Connectivity Extraction and Route Sequence Analysis) model, which leverages Graph Convolutional Network (GCN) to capture high-order dependencies, a hybrid route search algorithm combining Breadth-First Search (BFS) and Depth-First Search (DFS) to predict signal transmission routes, and a dual-stream dependency network to process route sequences. Extensive experiments conducted on large-scale real-world datasetsdemonstrate that HCE-RSA significantly outperforms several state-of-the-art methods, achieving superior prediction accuracy and effectively addressing the demand for personalized service recommendations.