The melting snow particles on top of clouds form the melting layer of precipitation. The melting process starts with the snow particles falling, so the microphysical characteristics of the melting layer vary continuously in vertical direction. In this paper, a Monte Carlo simulation model for the melting layer is developed, and the melting snow particles are modeled by more practical three-layered spherical particles. The size distribution of the melting snow particles is derived from the raindrops size distribution. Vertical profiles of radar reflectivity and specific attenuation factor are computed at 5, 10, 35 and 94 GHz using the Mie theory at rain rates below 12.5 mm/h. It is shown that the radar bright band can be absent in the melting layer at frequencies above 20 GHz. This agrees with radar observations at 35 and 94 GHz. Base on the radiative transfer theory, the Monte Carlo method is used to compute the reflectivity of the melting layer whose microphysical characteristics are continuous in vertical direction. We compared the reflectivity of the melting layers with two different size-distributions (Gamma size distribution and Marshall-Palmer size distribution). These provided theoretical and numerical basis for radar remote sensing of the melting layer with high frequencies electromagnetic waves.