The aim of this work is to improve the quality of water effluents coming from Elcogas IGCC power station (Puertollano, Spain) with the purpose of fulfilling future more demanding normative, using heterogeneous photocatalytic oxidation processes (UV/H2O2/TiO2 or ZnO). The efficiency of photocatalytic degradation for the different catalysts (TiO2 and ZnO) was determined from the analysis of the following parameters: cyanides, formates and ammonia content. In a first stage, the influence of two parameters (initial concentration of H2O2 and amount of catalyst) on the degradation kinetics of cyanides and formates was studied based on a factorial experimental design. pH was always kept in a value > 9.5 to avoid gaseous HCN formation. The degradation of cyanides and formates was found to follow pseudo-ftrst order kinetics. Experimental kinetic constants were fitted using neural networks (NNs). The mathematical model reproduces experimental data within 90% of confidence and allows the simulation of the process for any value of parameters in the experimental range studied. Moreover, a measure of the saliency of the input variables was made based upon the connection weights of the neural networks, allowing the analysis of the relative relevance of each variable with respect to the others. Results showed that the photocatalytic process was effective, being the degradation rate of cyanides about five times higher when compared to removal of formates. Finally, the effect of lowering pH on the degradation of formates was evaluated after complete cyanides destruction was reached (10 min of reaction). Under the optimum conditions (pH 5.2, [H2O2] = 40 g/l; [TiO2] = 2 g/l), 100% of cyanides and 92% of initial NH3 concentration are degraded after 10 min, whereas 35 min are needed to degrade 98% of formates. (c) 2006 Elsevier B.V. All rights reserved.