The control of pH is widely recognized as a difficult problem. A novel method for representing the nonlinearity of pH processes is developed. A system consisting of a combination of known and unknown chemical species is viewed as a combination of impulses (for the known species) and a piecewise continuous distribution (for the unknown species). This results in on-line identification of a small number of parameters that can be used to affect the process nonlinearity over a wider pH range than if only impulses were used. Computer simulations and experimental results using the proposed method for large step changes in concentrations of a simple chemical system demonstrate the performance of this approach. In particular, the simulations and experiments provide comparisons between the control of systems with and without identification and investigate the tuning of the controller and identification algorithm.