An induced gas flotation (IGF) machine is a type of water treatment equipment that separates oil droplets or particles from waste-water using the characteristics of floating air bubbles. In general, bubble movement inside an IGF is difficult to predict since the internal flow in an IGF is not linear due to vortex formation and recirculating flow. Furthermore, the actual bubbles are distributed in various sizes and it is very difficult to accurately simulate them. Therefore, the MUltiple-SIze-Group (MUSIG) method is used to correct this in the numerical analysis. The MUSIG method is to divide the bubbles into groups by size. It can predict relatively accurate bubble movements, coalescence and break-up. In this study, numerical analysis with the MUSIG model was conducted on various types of rotor blade configurations to determine water treatment. The results of void fraction, power consumption, air residence time, air mean bubble diameter, and drag coefficient were predicted and compared between the various configuration of rotor blade. Comparison of the results showed that the backward IGF rotor exhibited an increased void fraction by about 2.6 % and a decreased power number by about 37.6 % compared to the values seen using the reference model.