A monitoring concept for on-line sensors will be discussed which helps the WWTP staff to detect drift-, shift- and outlier effects as well as unsatisfactory calibration curves. The approach is based on the analysis of comparative measurements between the sensor and a reference method. It combines statistical analysis such as control charts and regression analysis with decision support rules. The combination of two different detection levels in the selected Shewhart control charts with additional criteria allows one to detect 'out-of-control' situations early with an optimized measurement effort. Beside the statistical analysis the concept supports the operator with a graphical analysis to monitor the accuracy of on-line measurements efficiently. The widely applicable monitoring concept will be illustrated with examples for an ion-sensitive NH4+- and a MLSS-sensor.