The demand for global freshwater is growing, while global freshwater available for human use is limited within a certain time and space. Its security has significant impacts on both the socio-economic system and ecological system. Recently, studies have focused on the urban water security system (UWSS) in terms of either water quantity or water quality. In this study, water resources, water environment, and water disaster issues in the UWSS were combined to establish an evaluation index system with system dynamics (SD) and geographic information systems (GIS). The GIS method performs qualitative analysis from the perspective of the spatial dimension; meanwhile, the SD method performs quantitative calculation about related water security problems from the perspective of the temporal dimension. We established a UWSS model for Guizhou province, China to analyze influencing factors, main driving factors, and system variation law, by using the SD method. We simulated the water security system from 2005 to 2025 under four scenarios (Guiyang scenario, Zunyi scenario, Bijie scenario and combined scenario). The results demonstrate that: (1) the severity of water security in cities is ranked as follows: three cities are secure in Guizhou province, four cities are in basic security and two cities are in a situation of insecurity from the spatial dimension of GIS through water security synthesis; and (2) the major driving factors of UWSS in Guizhou province include agricultural irrigation water demand, soil and water losses area, a ratio increase to the standard of water quality, and investment in environmental protection. A combined scenario is the best solution for UWSS by 2025 in Guizhou province under the four scenarios from the temporal dimension of SD. The results of this study provide a useful suggestion for the management of freshwater for the cities of Guizhou province in southwest China.