The data quality problems of OpenStreetMap(OSM) data, which is a typical kind of volunteered geographic information data, restrict the extraction of arterial road. Firstly, this paper presents a method to identify the morphological feature of arterial road through calculating road line density change rate and length of edges using Delaunay triangulation. Then the arterial road boundary identification model is established by integrating the two control conditions with different geometric dimensions. Finally, similar to forest fire spreading, a method of extraction arterial road polygon is proposed using Delaunay triangulation. An experiment based on OSM road data of 5 cities shows that this method achieves good and effective arterial road extraction results. Furthermore, a comparison analysis of roads attribute is also carried out to confirm the consistency between constructed road levels and experiment results.