Microcystins are potent cyanobacterial toxins which have been shown to be recalcitrant to conventional water treatment. Granular activated carbon (GAC) has been identified as a promising treatment option for the removal of these toxins. In this study, the adsorption of two microcystin analogues by GAC was described using experimentally-derived Freundlich and kinetic parameters. The Freundlich parameters (K and 1/n) were determined from bottle point isotherm tests, while kinetic parameters (Ds and kf) were determined using the short bed adsorber (SBA) test coupled with the homogenous surface diffusion model (HSDM). Once these parameters were derived, it was possible to predict the breakthrough of both microcystin analogues at a pilot-scale GAC filter. However, biological degradation, which was not modelled, appeared to be the predominant removal mechanism in the pilot-scale filter. Consequently, the HSDM overestimated microcystin breakthrough. A batch degradation experiment confirmed that bacteria within the biofilm of the pilot-scale GAC filter were capable of degrading the microcystins.