Abstract In order to confirm reliable particle size measurement and to prepare standard reference particles, experimental and theoretical studies have been conducted on the particle size measurement of three kinds of spherical glass beads. Revised-type sedimentation balance and microscopic methods with sample sizes greater than 10 000 were used for the measurement. The particle size distributions obtaine using the improved sedimentation balance and microscopic methods are nearly equal for the three kinds of glass beads. However, in the case of large particle diameters, particle sizes measured by the improved sedimentation balance method are larger than those obtained by the microscopic method. The reason for this is due to the fact that particles greater than about 50 μm settle too fast in a sedimentation bath. In the case of the microscopic method, variation of the evaluated mass median diameter decreases with the increase in sample size. The results of numerical simulation agreed with the analytical solution propose by Masuda et al. For the improved sedimentation balance method, it is desirable to disperse particles completely. At the beginning of measurement, it is recommended that a dispersing agent is pasted on the test particles directly and then the slurry solution is mixed in a dispersion bath with supersonic vibration.