For pollution research with regard to urban surface runoff, most sampling strategies to date have focused on differences in land usage. With single land-use sampling, total surface runoff pollution effect cannot be evaluated unless every land usage spot is monitored. Through a new sampling strategy known as mixed stormwater sampling for a street community at discharge outlet adjacent to river, this study assessed the total urban surface runoff pollution effect caused by a variety of land uses and the pollutants washed off from the rain pipe system in the Futian River watershed in Shenzhen City of China. The water quality monitoring indices were COD (chemical oxygen demand), TSS (total suspend solid), TP (total phosphorus), TN (total nitrogen) and BOD (biochemical oxygen demand). The sums of total pollution loads discharged into the river for the four indices of COD, TSS, TN, and TP over all seven rainfall events were very different. The mathematical model for simulating total pollution loads was established from discharge outlet mixed stormwater sampling of total pollution loads on the basis of four parameters: rainfall intensity, total land area, impervious land area, and pervious land area. In order to treat surface runoff pollution, the values of MFF30 (mass first flush ratio) and FF30 (first 30% of runoff volume) can be considered as split-flow control criteria to obtain more effective and economical design of structural BMPs (best management practices) facilities.