The Super Fast Breeder Reactor (Super FBR) utilizes supercritical light water as coolant, which changes from liquid-like high density state to gas-like low density state continuously in the core without phase change. In the preceding study (Noda et al., 2017), new concept of axially heterogeneous core with multi-axial fuel shuffling was proposed. The core consisted of two layers of mixed oxide (MOX) fuel and two layers of blanket fuel with depleted uranium (DU), which were arranged alternatively in the axial direction. The study showed that, with independent fuel shuffling in the upper part and lower part of the core, breeding performance could be improved by increasing the upper blanket fuel batch number while keeping the fuel batch number of the rest of the core unchanged, because of increased neutron flux in the upper blanket. However, the study did not consider influence of different coolant density histories in the different axial level of the core on the core neutronics. Hence, this study aims to reveal influence of the different coolant density histories through design and analyses of the multi-axial fuel shuffling core with two MOX layers and three blanket layers. The three levels correspond to the coolant density below, around, and above the pseudo-critical temperature. The neutronics calculations are carried out with SRAC 2006 code and JENDL-3.3 nuclear data library. Unit cell burnup calculations based on collision probability method are carried out for 5 different coolant density histories to consider influence of different neutron spectrum on breeding performance of the core. Influence of instantaneous coolant density changes on the core neutronics are considered by coupling core burnup calculations with thermal-hydraulics calculations based on single channel model. Influence of independent fuel shuffling of the upper blanket on the core neutronics (breeding performance and void reactivity characteristics) is investigated, followed by a similar investigation on the lower blanket. The differences between the two schemes are investigated since coolant density histories are greatly different between the upper blanket and the lower blanket.