Direct non-equilibrium molecular dynamics (NEMD) was used to simulate the thermal conductivities of the monolayer and the bilayer silicon functionalized graphenes along the length direction respectively, with the Tersoff potential and the Lennard-Jones potential, based on the velocity Verlet time stepping algorithm and the Fourier law. Simulation results indicate that the thermal conductivity of the monolayer silicon functionalized graphene decreases rapidly with increasing amount of silicon atoms. This phenomenon could be primarily attributed to the changes of graphene phonon modes, mean free path, and motion speed after silicon atoms are embedded in the graphene layer. Meanwhile, the thermal conductivity of the monolayer graphene is declined in the temperature range from 300 to 1000 K. As for the bilayer silicon functionalized graphene, its thermal conductivity increases as a few silicon atoms are inserted into the layer, but decreases when the number of silicon atoms reaches a certain value.