Abstract: Principles and operation of a sensor array for common water and wastewater analysis is described, together with fuzzy control of its operation such as pattern recognition for particular conditions of the water composition and the recognition of changes in the sensor performance. Field experiments were performed by a sensor array for conductivity, pH, and turbidity and a memory-programmable control for data acquisition. Stages of hard clustering by a main-component analysis (Henrion et al., 1986) and fuzzy clustering or alternatively, a neuronal network were used for data evaluation. The results indicated the successful recognition of 1% and 0.05% deviations of nonionic and anionic tensides, respectively. (SFS)