It is known that when water is a product in reactions catalysed by cation exchange resins, it inhibits the reaction rate much more than predicted by the reverse reaction or dilution effects. In this work the inhibiting effect is ascribed to the preferential association of the catalytically active sites with water. In the derivation of the kinetic model, a Freundlich type adsorption isotherm was used to quantify the number of sites occupied by water. This is combined with a power law expression for the reaction rate. The resultant expression can accurately predict the reaction rate for various initial concentrations of water and mesityl oxide. Even when water was initially added to the reaction mixture, this model still gave an absolute average error of 6.5% compared to a 54.6% error when the same approach was followed but with the more popular Langmuir isotherm to describe site deactivation. The kinetic expression previously proposed for this reaction system by Klein and Banchero also failed when water was added to the reaction mixture and gave an average error of 71.1%. The procedure used to derive the model is therefore suggested for all cation exchange catalysed reactions where water is one of the products.