Dear fiend Mircea Slanina, thank you for your attention and response !
You are right, my Saturn stackup model was wrong.
I looked on this as on Microstip, because it is Flexible PCB where the flexible area actually has only 2 layers, and it is major area. So, it seems, Hyperlynx doesn't support Rigid Flexible PCBs with different stackups over the same PCB. Or, just the exporting process from Altium to Hyperlynx doesn't support it.
But anyway, there is still significant mismatching between Saturn and Hypelynx diff impedance calculations .
In case you are showing, H2 layer is 5mil .
In my case , the H2 of Saturn means the dielectric row2 (DE_Top_Layer) in my stack-up, which is 25.2 mil .
Here is correct Saturn model, and I still expecting from Hyperlynx to get the same result . So, as before, I still missing here anything.