This study investigates the application of higher-order network models to urban systems, with a focus on the Sioux Falls transportation network as a case study. By simulating pathway data using MATSim and embedding this data into higher-order networks, we address the limitations of traditional first-order networks, particularly their inability to capture memory effects and complex pathway dynamics. Our findings demonstrate that higher-order networks offer a more precise representation of pathway patterns and outperform first-order networks in several key aspects. Experiments with path-fitted multi-order networks show enhanced performance in capturing betweenness centrality metrics, while networks constructed directly from observed pathways or firstorder topologies also exhibit superior topological features. This integrated approach, combining bottom-up simulations with top-down structural analysis, provides a robust framework for understanding urban infrastructure systems and offers deeper insights into the relationship between network topology and transportation behavior.