Mining tourists’ spatiotemporal patterns are fundamental for comprehending and applying spatial structure theories in tourism research. This study mines 27,981 tourist-generated trajectories in Beijing winter tourism to reveal the spatiotemporal patterns and spatial structures using a framework combining similarity measures and improved density-based spatial clustering of applications with noise algorithm. The findings indicate that Beijing winter tourism has established a spatial structure consisting of “three poles-two axes-two secondary poles.” A total of 50 popular routes and 12 patterns of winter tourism can be categorized into three main types and six subcategories: winter tourism leading (skiing oriented, skating oriented), winter tourism with sightseeing (urban snow scene sightseeing, mountain snow scene sightseeing), and winter tourism with leisure (skiing + hot spring therapy, skating + city walking). The findings demonstrate that identifying spatial structures through tourism pattern mining is scientific and feasible. This study examines the movement patterns of tourists to reveal tourism spatial structure and incorporates spatial geography theory to assure the accuracy and scientific validity of the identification process, which enables a comprehensive knowledge of the spatial structure of Beijing's winter tourism industry and broadens the application scenario of spatial geography theory in the era of big data.