Abstract: The effect of the particle spatial distribution on the thermal conductivity of composites and the essential condition of the formation of the effective thermal conductive pathways were investigated. In order to solve the modeling problem of the representative volume element (RVE) with any volume fraction and specified spatial configuration, the strategy to describe the objective spatial distribution configuration by the spatial distribution potential-energy function was employed, and a Monte Carlo controllable spatial distribution algorithm was designed, which can effectively create the RVE containing cluster and network configurations with any volume fraction. The simulated results show that, at the same volume fraction, the network configuration is easier to form the thermal conductive pathways and features higher thermal conductivity than the cluster one does; the volume fraction plays a key role in the formation of the effective thermal conductive pathways, which can occur only when the volume fraction is larger than 20% and the distance between the particles is short to some extent; with the increasing distance between the particles, the thermal conduction decreases in an exponent form. Therefore, a given amount of volume fraction and relative effective distribution of particles become two essential conditions of the formation of the effective thermal conductive pathways.