Influence of mass bias effect, isobaric and polyatomic interferences on the results of 234U/238U, 235U/238U, 236U/238U isotope ratio determination in uranium oxide by sector-field ICP-MS was studied. Uranium isotopic standards CRM U100, CRM U200 based on triuranium octoxide (U3O8) and single-collector mass spectrometer ICP-SFMS ELEMENT 2 were used for research. It has been demonstrated that the mass bias effect has most influence on the results of uranium isotope ratios determination. To investigate the influence of the mass bias effect on the determinations of uranium isotope ratios, the external standardization calibration was used with three models (linear, power, exponential) describing the behavior of the specific discrimination coefficient β versus the mass of measured isotopes. The mass discrimination factor has been found to vary from 6.00 ´ 10-3 to 1.20 ´ 10-2. The advantage for using the (power/exponential)-law models of the β=F(Δm) relationship for correcting measured isotope ratios was justified. In case of polyatomic interferences, the efficiency of uranium hydride ion (235U1H+) formation is 3.54 ´ 10-5, while the impact of isobaric overlapping due to the contribution of scattered 238U ions to the intensities of less abundant 236U and 235U ions reaches 8.17 ´ 10-6. The relative measurement error for the 234U/238U, 236U/238U , ratios was found to be < 0.5 %, and for the 235U/238U, ratio less than 0.1 %. The calculated standard uncertainty u of the 234U/238U, 235U/238U, 236U/238U isotope ratio measurements in the CRM U100 was 0.563, 0.322 and 0.856 %, respectively. These are reasonable estimates in comparison with the uncertainties of certified values of 0.296, 0.097 and 0.265 %.