<p indent="0mm">As one of the most used polymer material, polyethylene (PE) has excellent mechanical and processing properties. However, the low thermal conductivity limits the application of polyethylene in field of electronic devices, which needs good thermal conductivity for heat management. However, recent studies report excellent thermal conductivity for PE fibres. It should be very interesting to check if it is possible to convert PE from low thermal conductivity materials to high ones. In this study, in order to make a deep understand on thermal conductivity of PE material, a coarse-grained model of crystalline PE was constructed, and molecular dynamics simulation methods were used to explore the effect of crystallization on the intrinsic thermal conductivity of PE. First, Boltzmann inversion method was applied to build the coarse-grained model of PE, which is suitable to mimic physical behaviours of PE material with varied crystalline degrees. The simulation data such as the bond distribution function, angular distribution function, and radial distribution function of the coarse-grained model are found consistent with those from all-atom model. Furthermore, various physical data calculated from coarse-grained model are compared with those from reported experimental studies, such as glass transition temperature, thermal expansion coefficient, mean square radius of gyration, diffusion coefficient and thermal conductivity. All above physical data of coarse-grained model fit well with those from literatures. It means our coarse-grained model is readily to mimic PE materials, especially in field of thermal properties. For example, the curve of thermal conductivities of a single PE chain varied along with chain length calculated from our coarse-grained model lies between results from previous theoretic predictions. Second, the coarse-grained PE model with various crystalline degree and same crystalline degree of varied crystalline particles sizes are systematically investigated. We paid much attention on effect of stretching on thermal conductivity of PE models, because stretching is an efficient way to alter alignment state of PE chains. The results show that thermal conductivity of the PE model increases along with the increasing of crystalline degree. During stretching, the PE models with higher crystalline degree are easily reconfigured to align chains along the stretching direction, which is the thermal transporting direction, that in turn causes an obviously increasing for thermal conductivity. At the same crystalline degree, PE model with larger crystalline particle size is easier to form a high thermal conductivity system under stretching comparing with those with small crystalline particle sizes. Order parameters represent alignment of PE chains are calculated to analysize this observation. To conclude, the coarse-grained PE model established in this paper can be used to study the thermal conductivity of complex crystalline systems. The physical properties calculated from coarse-grained PE model are found to fit well with experimental and theoretic data of PE materials. Our simulation explores the dependence of thermal conductivity on crystalline state of PE material, and provides a theoretical basis for designing and manufacturing the intrinsic high thermal conductivity polymer materials. The coarse-grained model is thought to be able to explore mechanical properties of crystalline PE materials in the future.