云南省作为新一轮西部大开发的重要战略中心以及我国面向西南开放的重要桥头堡，面临经济社会快速发展机遇，同时又是我国西南地区重要粮食产区和林区生态安全屏障，建设用地需求增加和耕地保护矛盾日益突出。云南省低丘缓坡资源丰富，如何进行科学合理、可持续性强的低丘缓坡土地利用规划，实现土地利用可持续发展备受国内外关注。本研究将云南省的低丘缓坡土地利用规划分为三部分，第一部分是土地利用数量结构变化预测，第二部分是基于生态适宜性评价的用地类型分布区域限制研究，第三部分是土地利用空间布局优化，最终实现了从土地利用数量结构到空间布局的完整优化配置过程。本研究首先进行土地利用数量结构变化模拟预测，为土地利用空间布局优化提供基础。基于云南省2002年和2007年投入产出表、统计年鉴以及2000年和2005年遥感解译土地利用数据，构建了包含土地资源账户的社会核算矩阵（SAM矩阵），作为土地利用变化均衡分析模型（CGELUC模型）的数据基础。基于SAM矩阵，以区域内土地经济社会效益和生态环境效益最大化为目标，研究利用CGELUC模型模拟预测了基准情景、经济社会发展情景、生态环境保育情景下云南省2008-2020年的土地利用数量结构变化。结果表明三种情景下耕地、林地、未利用地和水域面积均逐年递减，且耕地面积减少最多，而林地和建设用地面积逐年递增，而经济社会发展情景下各种土地利用类型面积变化最大。同时，所有土地利用类型面积的年均变化量呈现先增加后减少的趋势。本研究在进行土地利用空间布局优化之前，开展了基于生态适宜性评价的用地类型分布区域限制研究，为土地利用空间布局优化提供可持续性的建议与指导。通过构建生态适宜性评价指标体系，利用层次分析法确定指标权重，采用加权指数和法得到云南省的耕地和建设用地生态适宜性评价综合得分，最后采用Natural Break法对适宜性进行等级划分，并综合考虑了特殊因子自然保护区的限制，将云南省划分为四个等级：不适宜区、限制区、中等适宜区和高等适宜区，从而得到用地类型分布限制区域，进一步提高了土地利用空间布局优化的合理性。最后，基于CGELUC模型模拟所得的土地利用数量结构变化结果以及生态适宜性评价的用地类型分布限制区域，本研究利用土地系统动态模拟系统（DLS模型）对云南省2008-2020年不同情景下的土地利用空间布局进行了精细栅格尺度上的空间模拟。结果表明经济社会发展情景下的各种用地类型的扩张与收缩最为剧烈与显著，建设用地大规模扩张，出现建设用地大规模集中连片的现象，而耕地和草地被建设用地大量侵占，大幅度收缩，出现集中连片减少的态势，建设用地扩张与耕地保护矛盾进一步加剧；而生态环境保育情景下各种用地类型的扩张与收缩最为缓和，建设用地持续扩张，但其扩张幅度较小，没有建设用地大规模集中连片分布的现象，耕地和草地面积的减少被控制在一个较为合理的水平上，没有出现集中连片减少的态势，林地面积也呈现均匀离散的空间扩张趋势，西南林区的生态安全屏障功能也得到了加强。本论文以生态适宜性评价所确定的用地类型分布限制区域作为约束条件，通过情景设定，开展了云南省多情景下的土地利用数量结构优化与空间布局优化研究，为云南省制定土地利用规划、土地资源管理、生态环境保护和经济社会可持续发展提供了重要的决策参考信息，也为全国的低丘缓坡土地利用规划提供重要指导。

As the important strategic center of new round of western development and important bridgehead to southwest, Yunnan province is faced with fast economic and social development, while Yunnan province is also one of the major agricultural regions and ecological safety shelter of China, thus, the contridiction between the increasing demand of built-up land and farmland protection has become increasing apparent. In this paper, mild slope land use planning in Yunnan included three parts: the first part was the prediciton of land use strutural change, the second part was ecological suitability assessment to get the restricted areas to ensure the sustainbility of land use spatial distribution, the third part was optimization of land use spatial distribution, which realized the complete optimal process from land use structure to spatial distribution.This paper firstly conducted the prediction of land use structure change in Yunnan, providing a base for the optimization of land use spatial distribution. An social accounting matrix including land resource account was consructed based on the 2002 and 2007 Input-Output table and Statistical Yearbook, 2000 and 2005 land use data from remote sensing interpretation, used as the data base of Computable General Equilibrium of Land Use Change Model. Aimed at maximizing the regional social-economic efficiency and ecological efficiency of land use, Computable General Equilibrium of Land Use Change Model was used to predict the land use structure change of Yunnan from 2008 to 2020 in the baseline scenario, socio-economic development scenario and environmental protection scenario. The result showed that the area of farmland, grassland, water area and unused land would decrease year by year, and farmland decreased the most, while the area of forest and built-up area would increase year by year. In socio-economic development scenario, the area change of all kinds of land use types were the most. Meanwhile, in all three scenarios, the annual change of all land use types would increase first, then decrease. Before the optimization of land use spatial distribution, this paper conducted a research of ecological suitability assessment and got the restricted areas to ensure the sustainbility of land use spatial distribution. We construted the index system for ecological suitability assessment, determined the index weight by analytic hierarchy process, calculated the ecological suitability score and classified Yunnan into four classes: unsuited area, restricted area, medium-suited area and high-suited area. Then we selected the unsuited area and natural reserve area as the restricted areas for land use spatial distribution to ensure the sustainbility and rationality of land use spatial distribution.Finally, based on the predicted result of land use structure change by Computable General Equilibrium of Land Use Change Model and restricted areas from ecological suitability assessment, this paper used the Dynamics of Land System Model to conduct land use spatial distribution simulation, which realized the spatial distribution optimization of different scenarios in Yunnan Province from 2008 to 2020 in a fine-grid scale. The result showed that in the socio-economic development scenario, the expansion and shrink of all land use types was the most remarkable, built-up land expanded a large scale and some districts showed large-scale concentration, while occupied by built-up land, farmland and grassland decreased sharply, which undoubtly aggravated the contradiction between expansion of built-up land and protection of farmland; contrary to the trend in the socio-economic development scenario, the expansion and shrink of all land use types was the most mitigatory in the environmental protection scenario, the built-up land showed a continuous expansion trend, while the expansion degree was much more small, thus, there was no phenomenon of large-scale expansion of built-up land, the shrink of farmland and grassland was controled in a ralatively reasonable scale, what’s more, forest showed an equally-discrete expansion trend, then the ecological security barrier function of southwest forest was also strengthened in the environmental protection scenario.With distribution-restricted areas concluded from the ecological suitability assessment as constraint condition, this paper designed three scenarios and conducted the land use structure change and spatial distribution multi-scenario researches of Yunnan, which provided significant decision-making reference information for land use planning, land resource management, environment protection and sustainable socio-economic development in Yunnan Province, and also provided important scientific basis for land use planning in mild slope areas.