use strict;

my @a = (1..20);
my @b = (2,3,13,14,17);

@a = do{my %i; undef @i{@a}; delete @i{@b}; sort {$a<=>$b} keys %i};
[download]

Another suggestion would be to get familiar with map which is a succinct way of writing a foreach. What map allows you to do is iterate over a list, be it an array, the keys of a hash, a range of numbers, letters, etc. and manipulate the contents element by element with a block of logic in a closure. In combination with grep map comes in very handy for just this sort of thing.

use strict;

#Lets say we have two arrays @a and @b
#and we want to, as in your posting, remove the elements in @b
#from those in @a

my @a = (1..20);
my @b = (5,8,13,16);

#we in essence re-write the contents of @a with our map statement
#that allows us to check the values for those we want to remove.
@a = map{
    my $i = $_; 
    (grep /^$i$/, @b) ? () : $i ;
    }@a;
[download]

The above code feeds in the existing elements of @a one at a time assigning each value to $_ within the closure/block {}, it then assigns this value to the locally scoped variable $i for reasons that will make sense in a second. grep is used to check for the value in our @b array, however since grep can also have a closure and it also assigns its values to $_ we check against $i (told you it would make sense). The logic for the check is contained in a ternary (if)?true:false; construct that returns either an empty list (), in effect nothing, or the value of $i if not found in @b, all of which gets fed into the new version of @a. The regexp anchors ^ and $ are used to keep partial matches from messing with things, like '2' matching against '12,20...29,32,42,52,etc'.
_{Seems like I have written this before... maybe because I have.}