To solve the problems encountered in the analysis of the large homogeneous and heterogeneous fast critical assemblies, Zero-Power Plutonium Reactor (ZPPR) 9, 10, and 13, we have revisited the analysis using improved methods. Two-dimensional cell calculations, cell calculations using multidrawer cell models, and three-dimensional transport theory core calculations were introduced. Using these methods, the discrepancies in the calculation-to-experiment (C/E) values of keff for the fast critical assemblies was reduced. The use of the multidrawer model reduced the C/E spatial dependency of the control rod worths in the ZPPR-10 cores. To investigate the remaining problems of the spatial dependence of the C/E values of reaction rate distribution and control rod worth, we have adjusted a cross-section set obtained from the JENDL-2 library using the integral experiments. The cross-section changes, particularly for the diffusion coefficient, 238U scattering and capture, and 239Pu fission cross sections, have corrected the spatial dependence, as well as the overestimation of the 238U capture to 239Pu fission rate ratio and sodium void worth.