Emergency evacuation in rural areas is more vulnerable than that in urban areas, requiring greater inter-village cooperation. The efficiency of inter-village evacuation cooperation is reduced when a road node or section near evacuation shelters with insufficient capacity becomes unusable due to congestion or damage. This paper used the evacuation road network in Xiji Town, Tongzhou district, Beijing, as a case study to supplement the existing theoretical system of rural emergency evacuation and improve rural evacuation efficiency. The topological statistical features of the rural road network were calculated using the complex network modeling method and Gephi. Changes in the size of the maximum connected subgraphs and edge count under deliberate and random attack strategies were analyzed using MATLAB. The results were then analyzed in conjunction with the spatial layout of rural emergency shelters for further analysis. The analysis showed that the rural evacuation road network in Xiji Town was sparse and poorly connected. Intersection numbers 30, 34, 35, 43, 46, 47, 79, 81, 82, 83, 86, 128, 130, 131, 149, 151, and 161 were critical to the road network's resistance to destruction. Therefore, rural shelters near these intersections, such as Laozhua, Fengge, Jinge, Yueshang, Qianzhai, Niumutun, and Xiyi, faced high risks when transferring excess evacuees. Finally, optimization strategies for vulnerable intersections around shelters with insufficient capacity were proposed. These strategies included optimizing layout and evacuation routes, improving transport circulation, enhancing intersection greening and landscaping, and strengthening the emergency functions of public spaces along roads.