It is very important to understand the temporal and spatial changes of residential land for the improvement of human settlement environment. Based on the three TM images of 1986, 1996 and 2005 and other assistant maps, the temporal and spatial characteristics of residential land in the Yuyang desert region of northern Shaanxi Province of China were analyzed by using geographical information system (GIS) and remote sensing (RS) technology. Some important conclusions were obtained, which mainly contained four aspects as follows. 1) The area of residential land in the Yuyang desert region increased 956.07 hm2 in total, and the changes of residential land among the townships were different with four various change trends. The expansion speeds in the Yuyang Township and Qinhe Township were the biggest from 1986 to 1996 and from 1996 to 2005, respectively. The expansion intensities in other townships were below 0.1%, except the Yuyang Township. 2) The number of residential land patches increased 2847, while the average area and perimeter of residential land patches decreased 0.304 hm2 and 60.78 m during 1986–2005, respectively. The residential land patches also displayed four different changes from 1986 to 2005 among the townships. 3) The residential land mainly distributed in the middle of Yuyang Township and Mahe Township, and the interface of Jinjitan Township and Mengjiawan Township. The total spatial pattern of residential land was relative stability, but the significant changes of residential land were found in the Yuyang Township and its vicinities during 1986–2005. 4) Three distribution patterns of residential land at the micro scale were categorized from the 100 typical villages in the study area. The number of villages in the pattern 1 increased, while the pattern 2 and pattern 3 decreased in recent 20 years, which implicated the adaptation of humans to the habitation environment. This research would provide scientific references for harmonizing the human-land relationship, optimizing the residential land spatial distribution patterns and constructing the well human settlement environment in the desert regions.