[en] Towards the practical use of fast reactors, it is a very important subject to improve prediction accuracy for neutronic properties in LMFBR cores from the viewpoint of improvements on plant economic efficiency with rationally high performance cores and that on reliability and safety margins. A distinct improvement on accuracy in nuclear core design has been accomplished by the development of adjusted nuclear library using the cross-section adjustment method, in which the results of critical experiments of JUPITER and so on are reflected. In the design of large LMFBR cores, however, it is important to accurately estimate not only neutronic characteristics, for example, reaction rate distribution and control rod worth but also burnup characteristics, for example, burnup reactivity loss, breeding ratio and so on. For this purpose, it is desired to improve prediction accuracy of burnup characteristics using the data widely obtained in actual core such as the experimental fast reactor 'JOYO'. The analysis of burnup characteristic is needed to effectively use burnup characteristics data in the actual cores based on the cross-section adjustment method. So far, a burnup sensitivity analysis code, SAGEP-BURN, has been developed and confirmed its effectiveness. However, there is a problem that analysis sequence become inefficient because of a big burden to users due to complexity of the theory of burnup sensitivity and limitation of the system. It is also desired to rearrange the system for future revision since it is becoming difficult to implement new functions in the existing large system. It is not sufficient to unify each computational component for the following reasons: the computational sequence may be changed for each item being analyzed or for purpose such as interpretation of physical meaning. Therefore, it is needed to systemize the current code for burnup sensitivity analysis with component blocks of functionality that can be divided or constructed on occasion. For this purpose, the burnup sensitivity analysis code has synthesized with an object-oriented scripting language. In the present study, an examination was conducted for the two-layer controlling model of the conventional system using Python, the object-oriented scripting language. On the basis of the result in the examination, a new analysis system for burnup sensitivity, PSAGEP (Python-wrapped SAGEP-burn), was implemented. It was confirmed the effectiveness of the reconstruction method based on the two-layer controlling model of conventional system. (author)