Classification of fine coal dust is very important in thermal power plant, in order to ensure complete combustion of fuel in the burner. Fine coal particulate is able to provide higher combustion efficiency as compared to coarse coal particulate, due to high surface area. In this investigation, modification of coal pulverizer classifier with new geometries were proposed, by extending the classifier blade into four different lengths: 4.36", 8.36", 12.36" and 16.36". Computational fluid dynamics (CFD) modelling was used to investigate the effects of classifier blade length to coal particles size distribution at outlet. CFD simulation of coal classifier model based on original design was initially done and validated by plant data. Next, four different blade length at a common fixed angle of 40°were tested, and the effects to resulting coal fineness were recorded. The simulation studies showed that the optimum blade length was found to be 12.36". The simulated classifier is able to achieve 71.5% of particles passing 75 μm sieve, with 57% classifier efficiency. A coal fineness improvement of 11% is achieved by the new design as compared to original design of the classifier model.