UO2-BeO is one of the most promising accident tolerant nuclear fuels due to its excellent thermal conductivity compared to pure UO2 fuel. Two different UO2-BeO fabrication methods have demonstrated the capability to fabricate enhanced thermal conductivity UO2-BeO composite fuels and improve fuel performance. In one method, BeO is continuously distributed around UO2 grains while BeO is dispersed in the UO2 phase in the other method. In the former type, BeO is considered as matrix and UO2 as dispersed particle. However, in the later type, BeO is considered as dispersed particle and UO2 as matrix. To calculate the thermal conductivity of UO2-BeO composite, Hasselman-Johnson model has been applied, which shows good agreement with experimental data. In this model, it includes the influence of thermal conductivity of matrix and particle, volume fraction of particle, radius of particle and the interfacial thermal conductivity between matrix and particle. To balance the improvement of thermal conductivity and enrichment of UO2, a UO2-BeO composite fuel with 10% volume fraction of BeO has been chosen. Besides, the grain size of matrix should be noticeable smaller than particle to relax thermal stress which may cause micro-cracks and destruction, leading to the grain sizes of UO2 in two types being distinctive, and resulting in very big effect on fission gas release. In the paper the thermal conductivity has been intensively studied as well as fuel performance in two different types of UO2-BeO fuels, and the two fabrication methods have also been compared to assess their applications in commercial reactors.