Uncertainty of roughness parameters has effect on soil moisture retrievals with backscatter models from Synthetic Aperture Radar observations. The uncertainty of soil moisture retrievals is important information for the usability of these estimates. In this paper we introduce a methodology to estimate the uncertainty of effective roughness parameters in the Integral Equation Method surface backscatter model, using a Bayesian Markov Chain Monte Carlo approach. Using Sentinel-1 imagery we demonstrate the methodology for a selected field, showing the posterior uncertainty distributions of the roughness parameters, and the effect on the backscatter model simulations and soil moisture inversions. The estimated total uncertainty of the soil moisture retrievals with the optimum parameter set is 0.043 m 3 /m 3 , which is slightly higher than the root mean square error of 0.040 m 3 /m 3 of the retrievals compared to in situ soil moisture measurements.