The security of drinking water resources is an important concern for human survival. The extensive use of geographic information system and remote sensing data in risk assessments of water quality from a spatiotemporal perspective can complement water resource management. In this study, we investigated Baimei reservoir conservation area, Fujian Province, China. The quantitatively inversed leaf area index was used to represent the vegetation growth condition in simulations of the hydrological process using soil and water assessment tool. According to the simulation results, the spatial distribution of the risk to water quality was mainly related to the land-use condition, and the temporal distribution of the risk to water quality was mainly related to the precipitation dynamics. Single factor analysis of various scenarios showed that uncontrolled expansion of agricultural land (scenario A) would lead to an overall average increase of 231.9% in the total nitrogen (TN) load in the basin. By contrast, if agricultural lands are all transformed into forests (scenario B), then the overall TN load in the basin would decrease by an average of 20.8%. In practical terms, conversion of sloping agricultural lands to forests (scenario C) could reduce the overall TN load by 17.0%. Thus, moderate conversion of agricultural lands to forests can reduce the pollution load effectively. The simulation results obtained for scenarios with adjustments to the fertilization time and amount showed that applying fertilizer on rainy days would increase the risk to water quality compared with days without rain. The method developed in this study can be applied for dynamic analysis in water quality risk assessments of water conservation areas and can provide a scientific reference for promoting ecological management strategies and sustainable development at the regional scale.