The Monte Carlo method for simulating photon migration in sample of farm products was introduced by means of analysis of interaction between light and sample of farm product, and a software of Monte Carlo simulation based on Windows was designed. The reflectance, mean optical pathlength and spatial sensitivity profile image were obtained by simulation with large numbers of photons. Moreover, Monte Carlo method, the most exact solution for light transport in tissue, was studied for modeling photon random migration in single-layered and multi-layered samples. The simulation result is in good agreement with experimental data on tissue phantoms. The results show that Monte Carlo method can provide theoretical predication in the design of near-infrared optic sensor.