Propagation of light into randomly dispersed scattering media is a complicated problem that can be numerically
simulated using statistical methods such as Monte Carlo. Jacques et al. have built three different Polarized Monte Carlo
(PMC) programs (under UNIX terminal) to keep track of the status of polarization of light traveling through scattering
media. However, several unexpected incidents such as running termination and datum overflow occur when the three
PMC programs are executed. In this paper, based on a detailed description of the operation procedure of the three PMC
programs, we analyze the reasons for termination and overflow and propose optimization method. The operation
procedure exhibits that our compulsory optimizations make the programs run normally. Comparison with previous works
shows that the optimized programs can generate more exact results.