Particulate flow is commonly encountered in industries as well as in many other chemical and mechanical engineering applications. The accurate measurement of particle size distribution is of the utmost importance since it decides the physical and chemical characteristic of the particles. The light extinction method can be used for in-line monitoring of particle systems thus providing real time measurements of both particle size distribution and particle concentration. In light extinction particle sizing, a classification inversion algorithm is proposed for the circular cylinder particles. The measured circular cylinder particle system is inversed with different particle distribution functions and classified according to the inversion errors in the dependent model. The simulation experiments illustrate that it is feasible to use the inversion errors of object functions to inverse the circular cylinder particle size distribution in the light extinction particle sizing technique. This classing inversion algorithm can avoid the defects that the type of the size distribution must be assumed beforehand for the light extinction method.