Under the changing environment of climate change and intensive human activities, flooding disasters occur at times. This paper presents a staged hydrological method based on Bai and Perron's (BP) multiple structural breaks test to provide a tool for analyzing hydrological data series amid such environment. This method is a unit root test of these breaks that combines structural mutation components with the traditional unit root test. For a data series, it does ordinary processing first, and then uses Bai and Perron's breaks test to check out the structural breaks, i.e. the start and end of flood season. Finally, these breaks are used to construct a corresponding staged data series. To demonstrate and verify the method, a BP test and traditional fuzzy statistical method are used to test the runoff series of the middle and lower Lancang River with the flood season from June through October. The results of these two methods agree well, showing that the staged hydrological method based on BP test is satisfactory but relatively more applicable to series of variable time scales.