Abstract: Under the background of increasingly blurred travelling and sightseeing, using GPS trajectory data to explore the spatial behaviors and temporal constraints of self-driving tourists in a large scale is of great significance for planning and designing tourist routes and tourism products. Taking the self-driving tourist hotspot - Tibet Autonomous Region as the study area, this paper collected 6748 footprints and 924 trajectories from Six Feet platform through Python, road data from Open Street Map (OSM), 540 Points of Interest (POI) from Baidu map. Methods like seasonal strength index, multi-ring buffer and kernel density were used to perform analysis on the spatio-temporal behaviors of self-driving tourists. Results show that: (1) tourists are mainly distributed along the traffic network in southern Tibet. The natural landscape around the national road has obvious appeal to self-driving tourists; (2) the number of tourists decrease in a power-function way as the distance from scenic spots increases, revealing that some tourists gather around the scenic spots. However, their distance is relatively large, and the landscape along the way has become an important tourist attraction; (3) during the year, influenced by the seasonal climate and natural environment, tourists travel mainly in summer and autumn around the developed cities, such as Lhasa and Linzhi. There are significant differences in driving speed in different seasons, and the relationship is autumn>summer>spring and winter; (4) during the day, under the constraints of eating, sleeping and other circadian rhythms and entertainment needs, tourists will stay at parking lots, service stations, roads with beautiful scenery and famous attractions. The overall rhythm of the tourism activities is a combination of moving and staying, and moving occupies most of the time.