Abstract A technique for the measurement of the residence time distribution (RTD) in microfluidic devices is presented. The measurements were performed by an input–response technique monitoring a dye tracer concentration spectroscopically at the inlet and outlet of a microfluidic device. The measurement setup ensures the interchangeability of microfluidic devices and thus allows characterising of many different devices containing diverse mixing structures and microchannel geometries. Since the experimental method is based on an external stimulus the measured RTD data show deviations from the actual RTD of the microfluidic device. These deviations are caused by the connected capillaries and their contribution to the overall RTD of the whole system. Therefore, modelling of the RTD is necessary. Two different models were applied in this study. The first one is the well-known axially dispersed plug flow model which is often used for macroscopic reactors. Since this model was developed for plug flow processes, its applicability to the laminar flow regime of microfluidic devices is strongly limited. Therefore, another empirical model was developed and applied in order to consider a wider range of different microstructures and process parameters. The RTDs of three different micromixers were investigated in a specified range of flow rates and modelled with the empirical model. Based on these results the mixing performance and integral flow behaviour of the different reactors could be analysed and discussed.