Abstract: Accelerated urbanization and high-intensity exploitation of resources have brought serious impact on the natural ecosystem and the quality of urbanization. Previous studies on this issue focus on grasslands, watersheds, wetlands, and mountains, while the ecological vulnerable karst regions receive little attention. At the same time, the selected indicators are based on statistical data which ignore the spatial differentiation characteristics, which is not favorable to regional comparative studies and lacks visualization effects. Guiyang, the capital city of Guizhou Province, located in a typical karst region, is an ideal area for this research, because ecological environments, rocky desertification and severe degradation characterize in this city. Besides, the city lacks arable land and has high population pressure and extensive pattern of economic development, forcing the urbanization process to accelerate and in turn increase the urban ecological security threat. This study is based on the theory of landscape ecology, using the "pressure-state-response (PSR)" model to build an urban landscape ecological security evaluation system of karst areas. The evaluation indexes include population pressure, urban sprawl, vegetation fractional coverage, the degree of rocky desertification, habitat quality index, urban heat-island effect and natural hazards. By comparing and analyzing the spatial distribution and evolution characteristics using multi-temporal of remote sensing data and GIS technology, this work obtains urban ecological security levels distribution in 2000 and 2013 for Guiyang City. The results show that the ecological security index of Guiyang City was relatively low in 2000 and 2013, with 3.04 more than the level of security in 2000 and 1.78 in 2013. In terms of spatial distribution, the security index tended to increase from southwest to northeast . The variation map of urban ecological security between 2000 to 2013 is obtained via superposition analysis by ArcGIS, in which the ecological security change area accounted for 27.89% of the total. The extremely strong degradation area is 0.15%, strong degradation area is 2.58% and the degradation area 20.10%; and the optimization area and strong optimization area are 4.55% and 0.46%, respectively. Moreover, classification of landscape types is done using the object-oriented method in the research. Next, this work overlays the classification results and urban ecological security data, and makes statistics to the values of ecological security at different landscape type. The results show that the ecological security of landscape types mainly concentrate at unsafe, critical, and secure levels. The urban areas mostly fall in a ill-conditioned level. With the increase of security levels, the proportion of occupying areas reduces gradually. The farmland landscape above in the safe level decreases and increases at other levels to some degree, but the overall trend is not obvious. The change trends of ecological security are roughly same in grasslands, forests and shrub landscapes. Wetland landscape type ecological security level is improved markedly. Comparing the ecological security of landscape types between 2000 and 2013, we find the ecological security indexes of all the landscape types tend to significantly decline except for wetlands. Degradation of farmland is more obvious and the wetland is the only improved landscape type. During the research period, landscape ecological security situation in Guiyang City indicates a trend of deteriorating. Therefore it suggests to optimize land use pattern in urban planning and construction in the future in order to raise the urban landscape ecological security level.