人口分布和增长对社会经济生态和环境有着深远的影响，获取不同尺度上高精度的人口空间分布信息对科学认识人口分布规律，正确制定人口政策，合理利用自然资源，快速评估重大灾害造成的伤亡面积等方面有重大意义。本文以2000年第五次人口普查数据为基础，基于土地利用数据，以GIS和SPSS软件为工具在陕西省内采取分区建模，分县控制，城乡分离的方法进行人口统计数据空间化,对数据空间化结果进行三维表达并定量分析人口分布与其影响因素关系。结果表明：本研究获得的1KM栅格人口精度高于全国范围1KM网格人口数据；陕西省人口集中分布在关中平原地区，关中平原地区占全省总人口的59.53%；估算1KM人口密度高值地区集中在关中平原西安市、渭南市、咸阳市和宝鸡市一带的城市群及其周边，其次为汉水谷地汉中市及其周边，人口密度最大值出现在西安市区（36206/平方公里）；人口密度低值地区集中分布在秦岭山脉、陕南大巴山地、陕北地区西部黄土高原区及北部沙漠地区。0-500米海拔地区分布全省45.04%的人口，年降水量500-600mm地区分布全省47.56%的人口，年平均气温12℃-14℃地区分布全省51.74%的人口，距离国道省道十公里范围内分布全省79.49%的人口。

It is widely recognized that population distribution and growth have a great impact on coordinated development of population, resources and environment. Only achieving high-precision information of population distribution, can we set up population policy correctly, make use of resources reasonably and evaluate the amount of disaster victims in time. Moreover, the population census data collected based on the administrative units do not match the environmental data collected based on the natural units. So studies on specialization of population statistics are necessary by modeling and estimating population data. The research scheme includes two parts: simulation and expression. In the part of simulation, effective research tools of Geographic Information System and Statistical Program for Social Sciences are used. Firstly, the study area is regionalized into three regions based on regional variation. Then, regression model based on Land use data is selected to simulate rural population. For the urban population, Clark model is transformed into new format based on the distance from the center of city and the area of city. Finally, the figure of simulated population of 1 km square grid-cells in Shaanxi is obtained. In the part of expression, three-dimensional map of population based on DEM in Shaanxi is made. The outcome of simulation shows that 59.53% of the population inhabit in the Mid-Shaanxi. In Shaanxi, population is centralized in metropolitan area of Xi’an, Xianyang, Weinan and Baoji. For the 21 townships in Shaanxi, which are selected to test and verify the model, these models are more efficient than Chinese 1km population data in the research area.