Oxidation and coagulation are two basic processes in removing organic pollutants from water. In this study, reaction rate constants and adsorption ability of twenty-eight organic compounds with different structures were implemented by Fenton experiments and Ferric-oxyhydroxides coagulation experiments. Based on the adsorption percentage, twenty-eight compounds were divided into two groups (group L, low adsorptive and group H, High adsorptive). Fifteen compounds with the adsorption percentage less than 15% were set as group L and the rest thirteen compounds with the adsorption percentage higher than 15% were set as group H. As for group L, the removal of fifteen compounds were mainly degraded by oxidation. While as for group H, the removal of thirteen compounds attributed to both oxidation and coagulation. According to the grouping, classified Quantitative Structure Activity Relationship (QSAR) models were subsequently developed by reaction rate constants and quantum chemical descriptors using multivariable linear regression. Results in this work showed that classification QSAR models exhibited different predictive ability for reaction rates. The classification models proposed the different mechanisms and characteristics of oxidation-dominated process and multi-process. The main factors governing reaction rate in oxidation process are q(CH+)(x), q(CH+)(n) and q(H+), they are relevant to the active site and hydroxide radical attack. While the frontier molecule orbital (E-gap (E-LUMO - E-HOMO) and E-sum (E-LUMO+ E-HOMO)) and active sites (q(CH+)(x) and f(0)(x)) play the important roles in multi-process. Furthermore, the quantum parameter analysis also provide a preliminary assessment of oxidation and coagulation in multiprocess.