The efficiency of daytime detection of flaming zones from space by thermal radiation is limited by the noise produced by solar radiation reflected in the "cloudy atmosphere - underlying surface" system. In this paper, the Monte Carlo method is applied to the study of statistical characteristics of solar and thermal radiation noise measured with a radiometer. The algorithm for numerical simulation is presented. The efficiency of the flaming zones detection depending on their size and cloud amount is estimated by the value of the signal-to-noise ratio. The mean thermal radiation noise is compared with the solar radiation noise.