Abstract: Large amounts of low and intermediate level radioactive waste (LILW) will be produced and piled up with the nuclear power industry development and the increasing widespread application of radioisotopes. The types of nuclides and the activity of the radioactive waste should be measured before the radioactive waste is disposed. Segmented gamma scanning (SGS) measure is an important technique of quantitative analysis for nuclear scrap and radioactive waste. However, when point source location changes, both the relative attenuation distance and scale efficiency are not corresponding, which causes errors of counts of detectors on activity estimation. This study is based on 12 different positions counts method and through Monte Carlo simulation methods, records waste barrels with uniform filling of polyethylene samples. It will find the max min counts records with Beer-Lambert attenuation law to value the turning radius, so that it can determine the location of radioactive nuclides in the current segment. Thus, the problem of the attenuation distance, scale efficiency non-correspondence of the radionuclide in different location which lead to SGS estimate error of point source activity is solved.