Sounds like the blue distribution has a higher mean value.
Univariate histogram plots will combine the curve colors when an X is in both distributions.
Do you not want that visual indicator ?
For example:
I used generated data with four classes to emphasize what the procedures do:
data have;
call streaminit(123);
do class = 0, 1;
do _n_ = 1 to 100+100*class;
x = rand('normal', 45 + 10*class, 4);
output;
end;
end;
run;
ods html file='univariate-normal.html';
proc univariate data=have;
class class;
var x;
histogram x / normal overlay;
run;
ods html close;

SGPLOT can also produce histogram plots, and overlay via the GROUP= option.
Review of produced images indicates the order of the histogram drawing is group mean descending (this is similar to the -mean being a z-index in CSS), so there is no way to force a higher mean distribution plot to be rendered wholly above of a lower mean distribution plot. The documentation does NOT discuss this situation, and there are no options for changing the group render order.

You might be able to roll your own plot by writing some custom code in GTL (graphics template language)