A comparative study on the energy dispersive X‐ray fluorescence analytical results of uranium determinations, in uranium and uranium–thorium mixed oxides, using specimens in the form of fused beads and pressed pellets, has been made. It was observed that in case of fusion bead specimens, the intensity of the analyte lines was approximately 1.6 times of that observed in pellet specimens under identical instrumental conditions. In case of uranium oxide samples, the analytical results with bead specimens were slightly better compared with the pellet specimens. However, in case of the uranium–thorium oxide mixtures, the average precision obtained with bead specimen was significantly better (1%, 1 s) in comparison with that achieved using pellet specimens (7%, 1 s). This difference may be due to the hardness of thorium oxide compared with uranium oxide, which affects the homogeneity of the pellet specimens prepared. In fusion bead method of sample preparation, even highly refractory material like ThO2 forms uniform glass beads. Addition of internal standard further improves the analytical results, with reduction in the percent deviation of energy dispersive X‐ray fluorescence results from the expected values to 3% from 7% compared with that obtained using without internal standard. The fusion bead method of sample preparation will be very useful for characterization of sintered (U,Th)O2 pellets, which are highly refractory and difficult to dissolve. Copyright © 2016 John Wiley & Sons, Ltd.