Abstract The ex-core detector response calculation is an important part in reactor design. However, the response function cannot be measured by experiments quantitatively. Ex-core detector response simulation is therefore required. For decades, the S N code has been used as the dedicated tool. Nowadays, more and more engineers are expressing an interest in using the Monte-Carlo method instead of the S N method in simulations, as it is expected that the Monte-Carlo method will give higher accuracy. In this paper, the modeling and simulation of ex-core detector responses is briefly reviewed based on the Korean Kori Unit 1 reactor. Then, the differences between the S N simulation and Monte-Carlo simulation are compared. The sensitivity of computational conditions is also discussed. It is shown that the problem dependence of cross sections and meshing dependence of spatial discretization in the ex-core detector response calculations are not as strong as expected. However, the ray effect is the main shortcoming for the S N calculation. Based on the analysis, two benefits are shown by using MCNP for the direct 3D calculation. Firstly, the impact of ray effect is eliminated without using the S N angular discretization. Secondly, the direct 3D calculation is easier to perform based on the powerful ability of 3D modeling and parallel computing of the Monte-Carlo code. The new DRF values are adopted in the dynamic control rod reactivity measurement of Kori Unit 1 reactor. The results show that the new DRF values improve the error of measured control rod worth by a percentage of 3.