The goodness-of-fit test is often used to select the clutter distribution model, but it cannot solve the problems of overfitting and the multi-model optimization. In this paper, a new method based on Bayesian Information Criterion (BIC) is proposed for clutter model selection, and the Reverberation Chamber (RC) is taken as the simulator of the radar clutter environment. The electromagnetic environment with a time domain fluctuation of the electric field the same as the clutter distribution is synthesized in the cavity by changing the settings of RC. The measured data is analyzed by using KS test and BIC. The results show the BIC method makes up for the deficiency of goodness-of-fit tests, and can be used to select the optimal distribution from several candidate distributions. In under-moded (low frequency) conditions of the RC, a clutter with Weibull distribution is synthesized in the RC;in over-moded (high frequency) conditions of the RC, a clutter with Rayleigh distribution will be synthesized.