In this study, the spectral interferences are investigated for the analytical peaks at 63.3 keV of Th-234 and 1001.0 keV of Pa-234m, which are often used in the measurement of U-238 activity by the gamma-ray spectrometry. The correction methods are suggested to estimate the net peak areas of the gamma-rays overlapping the analytical peaks, due to the contribution of Th-232 that may not be negligible in materials rich in natural thorium. The activity results for the certified reference materials (CRMs) containing U and Th were measured with a well type Ge detector. The self-absorption and true coincidence-summing (TCS) effects were also taken into account in the measurements. It is found that ignoring the contributions of the interference gamma-rays of Th-232 and U-235 to the mixed peak at 63.3 keV of Th-234 (U-238) leads to the remarkably large systematic influence of 0.8-122% in the measured U-238 activity, but in case of ignoring the contribution of Th-232 via the interference gamma-ray at 1000.7 keV of Ac-228 to the mixed peak at 1001 keV of Pa-234m (U-238) results in relatively smaller systematic influence of 0.05-3%, depending on thorium contents in the samples. The present results showed that the necessary correction for the spectral interferences besides self-absorption and TCS effects is also very important to obtain more accurate U-238 activity results. Additionally, if one ignores the contribution of Th-232 to both U-238 and K-40 activities in materials, the maximum systematic influence on the effective radiation dose is estimated to be similar to 6% and similar to 1% via the analytical peaks at 63.3 and 1001 keV for measurement of the U-238 activity, respectively. (C) 2009 Elsevier Ltd. All rights reserved.