Introduction: Assessment of radiation DOSE is an important task in radiation protection dosimetry whereas ABSORBED and effective DOSE measurements directly in body organs are impossible. So Monte Carlo simulations is necessary to estimate radiation DOSE. The method of DOSE calculation, the body model and the computational code can be mentioned as three main factors that have an affect on dosimetric quantities. The aim of this study is the determination of the above factors influence on the ABSORBED DOSE and effective DOSE evaluation. For this purpose different comparisons between the mathematical and the voxel phantoms were done.Methods: ORNL modified adult phantom is the mathematical phantom which is described by Cristy and Eckerman in 1987. This phantom is utilized for calculation of the ABSORBED DOSE. The effective DOSE is calculated according to the manner which is introduced in the ICRP reported 60 and 103.Results: All of the DOSEs Evaluation have an uncertainty less than 0.5%. MCNPX code and ORNL modified adult phantom are applied for the DOSE assessment in the energy range 10-9 – 20 MeV, under AP, PA, RLAT, LLAT, ROT and ISo irradiation conditions. The results are then normalized to the unit of neutron fluence. The calculated ABSORBED DOSE was compared with ICRP74 data in 9 organs and with Bozkurt et al data in 18 organs. The effective DOSE was calculated for whole body. Then these data were compared with results of a mathematical phantom and some voxel models in different irradiation geometries.Conclusion: Although the ABSORBED DOSE results of ORNL show fewer differences with ICRP data than Bozkurt et al data, one can deduce neither complete agreement nor disparity between the ORNL data and the other two data sets. Totally 73% of the ORNL data in comparison with MIRD data and 48% of data in comparison with VIP_MAN results have difference less than 15%. Comparison of the ORNL effective DOSE with some male and female voxel phantoms (TARA, HANAKO, GSF, VIP_MAN) and an analytical model (MIRD) show that the changing of body model and computational code have few influence (less than 15%) on the effective DOSE results. But changing of wR and wT parameters have a significant effect on the results so that the largest discrepancies are about 100% on some data.