With material science demands on ever smaller particle sizes, new approaches and effective methods of their classification are needed. Turbulent flow field patterns and particle trajectories inside of a dynamic air classifier are investigated. Increasing computing power together with new turbulence models and approaches to simulate complex fully turbulent problems by solving Navier-Stokes equations allows studying and capturing smaller flow structures and properties more accurately. Particles below 10 microns are more susceptible to smaller local vortexes and particle fates are therefore more dependent on these local structures. Area of focus are the conditions required for classification of particles with sizes below 10 microns as this size is at the limit of air classification method possibilities. CFD software and the latest knowledge in turbulence modelling are used to numerically simulate flow field inside a dynamic air classifier. Effects of varying operating parameters on flow patterns and discrete phase classification outputs are investigated. Experimental verification of the simulated flow fields includes advanced imaging method (PIV) measurement of flow velocity and is used to visualize flow field structures in the classifier rotor blade passageway region. Predicted particle trajectories and their fates are experimentally verified by classification trials carried out on dynamic air classifier and the particle distribution curves are established by laser diffraction method. Tromp curves and efficiency of classification process are studied.