This paper investigates the effectiveness of granular activated carbon (GAC) for removing off-flavor compounds of water-geosmin and 2-methylisoborneol (MIB). More specifically, it focuses on the development of an appropriate modeling approach and experimental protocol for the design of fixed-bed GAC adsorbers. Adsorption equilibrium, rate and long-term mini-column studies were conducted on a bench-scale for geosmin and MIB to estimate the equilibrium and mass-transfer parameters required for adsorber modeling. The dispersed flow homogeneous surface diffusion model (DFHSDM) was used for the prediction/simulation of the adsorber dynamics. Scale-up procedures based on dimensional analysis and similitude were employed for the design of full-scale adsorbers from bench-scale adsorbers, and for performance forecasting of full-scale adsorbers under different operating conditions. More importantly, operation and maintenance (O & M) costs were estimated for full-scale adsorbers directed at the removal of off-flavor compounds, based on carbon utilization rates and disposal costs. These estimates were obtained for different plant capacities and empty bed contact times (EBCTs).