Abstract: The thermal property of AlN filled silicone encapsulant was investigated through experimental method and numerical simulation. A 3D finite element model of AlN filled silicone encapsulant was built by random sequential adsorption (RSA) method. This model can generate representative volume element of random particle location and different particle diameters. The single particle size and multiple particle sizes models can be established. The AlN particles in 3 μm filled silicone encapsulant composite was prepared and the thermal conductivity was tested at different filling volume. The model was verified to be accurate by comparing the predicted and the practical thermal conductivity. The simulation results show that thermal conductivity of the silicone encapsulant increases with the filler mass fraction, and the thermal conductivity increases rapidly after the filler mass fraction being higher than 50%. When the AlN particles with normal distribution pack effectively, the filling volume is increased. The thermal conductivity is different when the composite is filled with different mass ratio of two different particles.