Methods for calculating the physical properties (thermal conductivity) of composite materials (CS) based on empirical and numerical models are investigated. These use the idea of Maxwell and Rayleigh about the proportionality of this parameter to the physical properties of the matrix and filler, as well as the concentration of the filler in the matrix. These models can consider the shape of the particles and their relative position. Comparison experimental data showed that such models are not universal and work only in the field of low concentrations of filler in the matrix. Another class of effective thermal conductivity models allows you to consider Thermal boundary resistance of the contact between the filler particles and the matrix (Kapitsa thermal resistance). In this case, there is a serious problem of determining the thickness of the interfacial thermal layer, which can only be estimated by indirect methods. In general, the effective thermal conductivity of the composite (ETC) is also influenced by other factors: contact thermal resistance between filler particles, aggregation of filler particles, dependence of the physical properties of the filler on particle size, ordering of matrix molecules in the interfacial region. Numerical calculation is the most efficient because it is possible to consider the influence of a lot of factors, which is not possible in the case of an empirical model. A composite is a heterogeneous structure, so the generalized heat flux method is applicable. The average heat flux through the composite is calculated and then the effective thermal conductivity coefficient of the composite is calculated. The calculation of ETC is performed using the generalized heat flux method (GHFM). To solve the transient thermal conductivity equation, the Alternating Direction Implicit Method (ADI) was used. Methods for estimating the thickness of the Thermal boundary are proposed based on comparison of experimental data and the results of a computational experiment and considering the dimensional factor. The concentrations of the filler at which these factors have the greatest effect on the thermal conductivity of the composite have been determined. Software has been developed by applying the methods described.