Dissolved air flotation (DAF) is a method for removing particles from water using micro bubbles instead of settlement. The process has proved to be successful and, since the 1960s, accepted as an alternative to the conventional sedimentation process for water and wastewater treatment. However, limited research into the process, especially the fundamental characteristics of bubbles and particles, has been carried out. The single collector collision model is not capable of determining the effects of particular characteristics, such as the size and surface charge of bubbles and particles. Han has published a set of modeling results after calculating the collision efficiency between bubbles and particles by trajectory analysis. His major conclusion was that collision efficiency is maximum when the bubbles and particles are nearly the same size but have opposite charge. However, experimental verification of this conclusion has not been carried out yet. This paper describes a new method for measuring the size of particles and bubbles developed using computational image analysis. DAF efficiency is influenced by the effect of the recycle ratio on various average floc sizes. The larger the recycle ratio, the higher the DAF efficiency at the same pressure and particle size. The treatment efficiency is also affected by the saturation pressure, because the bubble size and bubble volume concentration are controlled by the pressure. The highest efficiency is obtained when the floc size is larger than the bubble size. These results, namely that the highest collision efficiency occurs when the particles and bubbles are about the same size, are more in accordance with the trajectory model than with the white water collector model, which implies that the larger the particles, the higher is the collision efficiency.