Abstract: Abstract: Analyze the forest fire emergency road network of Liangshui National Nature Reserve and optimize the route selection to meet the road network density requirements of 2025 in the national forest fire prevention plan (2016-2025). K-means clustering algorithm is used to select road nodes. Considering terrain and other environmental factors, ArcGIS multi factor superposition analysis is used to select the overall line. The line is verified to be feasible through field investigation. Establish an evaluation system for emergency roads in forest areas. Finally, the forest road density will be increased to 3.22 m/hm2, which is in line with the national expectation of forest road density in 2025. Regional area mean in regional segmentation index (S`)and regional area variance (D) decreased by 45.7% and 94.7% respectively, which provides a reference for the establishment of emergency road network in Liangshui protection area.

Key words: Key words: forest road network density, analytic hierarchy process, K-means clustering, GIS route selection figure, forest fire emergency road network