Based on the successful predictions of transient drop sizes in the first part of this research (Maaß et al., Prediction of drop sizes for liquid–liquid systems in stirred slim reactors—part I: single stage impellers. Chem. Eng. J., 162 (2010) 792–801), this part is a straight continuation and extension of the earlier work. The predictive capabilities of the used population balance equation model are increased for single stage impellers and transferred to scale-up procedures of such applications. Therefore different scale-up rules for liquid–liquid systems are tested experimentally and by simulations in two different sized, geometrically similar vessels. The multi stage impellers are tested against comparable single stage impellers in terms of power consumption, mixing time and minimum impeller speed. Especially for high aspect ratios (larger than three), multi stage impellers successfully compete with the single stage ones. The measured drop size distributions in slim reactors with multi stage impellers showed no dependency on the local position, although the dispersion process is tedious due to the compartmentalization. The simulations are not able to reflect this initial phase of the dispersion process, but are in close agreement with the experiments after complete dispersion is fulfilled. Based on these experiences the aspect ratio is increased up to five and the resulting drop size can be predicted with reasonable deviations (lower than 10%). The results of the scale-up of this multi stage impeller liquid–liquid system do not lead to a clear conclusion. Although the simulations recommend the use of constant power input, the experiments could not support this. None of the other traditional scale-up rule are supported by the experiments. Overall, the results of power consumption, mixing time and dispersion behavior show the great potential of multi stage impellers for process optimization and intensification in slim reactors.