The use of ultraviolet (UV) disinfection in water treatment is governed by several factors, including flow field, fluence rate field, and microbial inactivation kinetics. In this study, a computational fluid dynamics (CFD) model was developed for UV disinfection in a closed-conduit reactor where an improved low-Reynolds number k-epsilon model was used to calculate flow field and a modified P-1 model was employed to obtain the fluence rate field. The Chick-Watson model was adopted to characterize the inactivation of microorganisms. Commercial CFD software FLUENT 6.3 was employed to solve the governing equations. The predicted flow field agreed well with experimental data obtained by digital particle image velocimetry (DPIV) (Liu eL al., 2007) in terms of velocity field. The proposed CFD model was also evaluated by comparing current predictions to bioassay test data, and reasonable agreement was obtained in terms of effluent log inactivation. The impact of wall reflection of the light on the fluence rate field and the viable microorganism concentration field was investigated. The effect of wall reflection of the light on effluent log inactivation was also investigated under different water qualities and lamp power conditions. The results showed that at higher inactivation levels, the effect of wall reflection was more influential. (C) 2011 Elsevier Ltd. All rights reserved.