Since numerical simulation has become a popular method for studying the effects of land use and land cover change on climate and environment, spatially explicit historical cropland datasets are increasingly required in regional and global climate change and carbon cycle research. In this study, using historical population data as a proxy, we estimated the provincial cropland area of Qinghai and Tibet in 1910. Based on the statistical data of the National Bureau of Statistics of China, the survey data of the Ministry of Land and Resources, and the results of some previous studies, we revised the cropland area of Qinghai and Tibet in 1950-2000. The relationship between altitude and surface slope and cropland distribution were quantified to develop the spatially explicit reconstruction model of historical cropland at a resolution of 1 km*1 km. Since the cropland area reached the maximum in the 1980s, the satellite-observed cropland distribution extent of this time period was taken as the maximum distribution extent of historical cropland. The model developed in this research was used to reconstruct the spatial patterns of cropland in Qinghai and Tibet in 1910, 1960, 1980, and 2000. The reconstruction results show that: (1) in 1910-1950, cropland area of Qinghai-Tibet was stable, while in 1950-1980 cropland area increased rapidly, reaching 10583 km2, which is the maximum of the entire study period; in 1980-1990, cropland area decreased slightly; and in 1990-2000, cropland area increased slightly; (2) with regard to its spatial distribution, in 1910-1960, cropland expanded and land use activities intensified greatly in the Yellow River-Huangshui River Valley (YHV); in 1960-1980, cropland expansion and land use intensification occurred in the YHV, the Yarlung Zangbo River, the Nianchu River, and the Lhasa River valleys; in 1980-2000, the spatial pattern of cropland in Qinghai and Tibet remained unchanged. By comparing the reconstruction results of this study for 2000 with satellite- observed cropland distribution of the same year, we found that the correlation coefficient was 0.92 and the absolute difference followed normal distribution. The percentage of grid cells where the absolute difference is low (-10% to 10%) reached 73.29%, while the percentage of grid cells where the absolute difference is high (>40% or <-40%) was 1.94%. Incorporating more information on historical population and cropland of Qinghai and Tibet will help improve the accuracy of our reconstruction modeling. The reconstruction results of this research can be used in regional climate models to study the impact of cropland cover change on the climate and carbon cycle.