Abstract: Based on a kind of radioactive aerosol spectrum monitor, the effect of importance sampling to Monte Carlo computing efficiency was analyzed in the paper. In order to transform the sampling method of polar angle from uniform sampling to logarithm sampling, one bias parameter was introduced in the analog simulation, and the weighted value of particle was adjusted finally to obtain correct results. In this way, the importance sampling method was implemented in the software. The importance sampling method was fully implemented in Geant4, and the effect of bias parameter on the computing efficiency was analyzed too. Simulation results indicated that the efficiency of simulation was largely improved by the importance sampling method, and the value of bias parameter should not be too large in practice.