In terahertz (THz) scanning imaging system, the image resolution is constrained by the size of beam spot, while the image quality is greatly affected by the fluctuation of the laser output power. Considering the prior information, Bayesian rule is applied to improve the confidence of image processing based on Markov random field (MRF) and simulated annealing (SA) algorithm. An MRF model is applied to denoise the simulated THz scanning image. Simulations are carried out to study the effect on the denoising result from the parameters such as Gaussian noise, cooling times, and inner iterative times. The effect of the denoising method is evaluated by observing the signal to noise ratio (SNR) and the root mean square error (RMSE). The relationships among cooling times, inner iterative times, SNR and RMSE are obtained. The simulation results show that the MRF method can effectively suppress the Gaussian noise while preserving the image details as well.