Particle size and particle size distribution are measured in-line in fluidized bed processes. Coating processes demand detection of the particle growth and the timely recognition of undesirable agglomeration for process intervention and process parameter change. The aim was the detection of agglomerates (large particles) in lots of small particles. Therefore spiking trials with pellets of microcrystalline cellulose were performed. The blends were fluidized in a fluidized bed apparatus and particle size was measured in-line by Spatial Filter Velocimetry method. Snapshots of volume density distribution and number density distribution at distinct moments of the process were registered, and finally the density distributions were averaged over the whole process time. Agglomeration was unequivocal detected by volume density distribution. In contrast the number density distribution was not suitable to detect low absolute numbers of large particles in the blends. In real fluidized bed coating processes with pellets and model substances, median particle size values x(10,3); x(50.3) and x(90,3) were measured versus process time. Median values increase continuously with process time under optimized process conditions. Due to worst case conditions, caused by super-wetting of the fluidized bed, Agglomerates were generated and detected by a non-linear increase of the x(90,3)-time-plot (greater storage of large particles). Also snapshots of volume density distribution under worst case process conditions indicate clearly the formation of agglomerates. The in-line particle size measurement at fluidized bed coating is a useful Process Analytical Tool (PAT) in the process control. (C) 2016 Elsevier B.V. All rights reserved.