I may be missing something... but wouldn't using the Perl built-in substr, i.e. a simple substr($seq, -20) do the job?

(A negative offset returns that many chars counting from the end of the string.)

Actually you don't need a bioperl module for this.
You can use the builtin substr function instead:

#!/usr/bin/perl -w
use Bio::DB::GenBank;

$db_obj = Bio::DB::GenBank->new;

$seq_obj = $db_obj->get_Seq_by_acc(EU864037);

$sequence = $seq_obj->seq;
$pos = length($sequence)-20;            #go to position from end
$seqend = substr($sequence,$pos,20);

print "whole sequence: $sequence\n\n",
       "Just the end: $seqend \n"; #just for verifying
[download]

    # gets sequence as a string from sequence object
    $seqstr   = $seqobj->seq(); # actual sequence as a string
    $seqstr   = $seqobj->subseq(10,50); # slice in biological coordina
+tes
$seqobj->length()            # length
[download]

$segstr = $segobj->subseg( $segobj->length - 20, $segobj->length );
[download]

Thanks dude....