This study was conducted to find end-members and tracers which are effective to be applied in the End Member Mixing Analysis (EMMA) model for runoff separation at the Gwangneung coniferous forest catchment (13.6ha), Gyeonggido, Korea. We monitored three successive rainfall events during two weeks from June 26, 2005 to July 10, 2005, and analysed chemical properties of rainfall, throughfall, stemflow, groundwater and soil water considered as main components of storm runoff. The followings are the results of analyses of each component and tracer. Groundwater, soil water and rainfall (or throughfall) were dominant runoff components. Rainfall and groundwater were selected as main components for the two components-one tracer mixing model, and groundwater, soilwater and throughfall were selected as main components for the three components-two tracers mixing model. Tracers were selected from anion ($Cl^-$ and ${SO_4}^{2-}$), cation ($Na^+$, $K^+$, $Mg^{2+}$, and $Ca^{2+}$) and Acid Neutralizing Capacity (ANC) in event 1, 2, and 3. $Na^+$, $Ca^{2+}$ and ANC were selected in the two components-one tracer mixing model and ${SO_4}^{2-}-K^+$, ${SO_4}^{2-}-Na^+$, ${SO_4}^{2-}-Ca^{2+}$, ${SO_4}^{2-}$-ANC, and $Ca^{2+}$-ANC were selected in the three components-two tracers mixing model. Selected main runoff components and tracers can provide basic information to determine the contribution rate of each runoff component and identify the runoff process in a forest watershed.