Natural organic matter (NOM) is ubiquitous in the aquatic environment and is of concern because it impacts engineered and natural processes. For example, during water treatment, NOM reacts with chlorine to form disinfection byproducts that may be associated with adverse health effects. The conventional water treatment process sequence of coagulation, flocculation, clarification, and filtration is the most common treatment approach used to remove NOM. Anion exchange treatment, however, has the potential to be more effective than conventional treatment. To realize the potential of anion exchange technology, an improved understanding of the interactions among NOM, raw water characteristics, anion exchange resins, and process operating parameters is required. The experimental scale, flow regime, and test water were systematically varied to investigate the removal of NOM by anion exchange treatment. Bench-scale batch experiments were conducted using synthetic model waters containing NOM isolates and commercially available anion exchange resins, including a magnetic ion exchange (MIEX) resin. The charge density of the NOM isolates was fundamental to understanding anion exchange reactions. Ion exchange was clearly shown to be the mode of removal of NOM by anion exchange resins. Pilot-scale continuous-flow experiments were conducted using a local raw drinking water and MIEX resin. The effective resin dose, which is the product of the resin concentration and the resin regeneration ratio, was identified as the most important process operating parameter. The pilot plant study demonstrated that anion exchange treatment could be operated more effectively than previously believed. In addition, the pilot-scale continuous-flow tests and bench-scale batch experiments gave consistent results. A mathematical model describing the removal of NOM by anion exchange in a completely mixed flow reactor was developed based upon insights gained from the pilot-scale continuous-flow study. Model predictions were found to be in good agreement with experimental data. The validated mathematical model was used to evaluate the relative influence of operating parameters, anion exchange resin properties, and NOM characteristics on process performance. The mathematical model was also used to evaluate various treatment scenarios. This dissertation presents a unified framework for understanding the removal of NOM by anion exchange treatment. Advisors/Committee Members: Boyer, Treavor, Singer, Philip, Aiken, George, Christman, Russell, DiGiano, Francis A., Jorgenson, James, Miller, Cass T..