Green infrastructure plays an important role in reducing stormwater runoff and improve water quality. However, the influence of the proportion and pattern of green infrastructure on runoff reduction and water quality improvement has been an unexplored issue. In this thesis, we use SIMMAP model to simulate stochastic green infrastructure patterns, and calculate pattern indices. Furthermore, we import those patterns into SWAT model to simulate the rainfall-runoff process, and calculate the hydrological and water quality outcomes. We conducted regression analysis between the marginal effectiveness (such as runoff reduction, organic nitrogen, and phosphorus improvement) and the green infrastructure ratio. As a result, we found that 30% might be an optimal proportion for green infrastructure planning. Besides, the correlation analysis outcome between hydrological & water quality indices and pattern indices implies that larger sized, more convoluted, and more aggregated green infrastructure patterns have higher effectiveness to mitigate runoff risks and depollute water.