It is something like this:

!#/usr/bin/perl -w
use strict;

use Net::FTP;

my $hostname = "sample-host";
my $ftp = Net::FTP->new($hostname);

$ftp->cwd("mypath");

my @cat1 $ftp->ls("h?_*");
my @cat2 $ftp->ls("*-zips");

print "category 1:\n", join "\n",@cat1;
print "category 2:\n", join "\n",@cat2;
<>;
[download]

I think I must prefer globbing than regex since not all user that will use my program has knowledge in regex.

I have tried converting all * and ? into .* and .? respectively and use regex but before I finish coding it I thought about the "+" character which is a valid char for naming file or directory. It will be a problem when a file/folder contains a "+" char. since it is a special char in regex. And there are many other special char in regex which i don't want to make a code just to escape them all.

It seems that it will become complicated if i use regex. What should I do..?

which i don't want to make a code just to escape them all.

Why not? It's just a question of using quotemeta($1) instead of $1.

my $re_pat = '';
for ($glob_pat) {
    /\G ([^?*\\]+) /xsgc && $re_pat .= quotemeta($1);

    /\G \?    /xsgc && do { $re_pat .= '.';           redo };
    /\G \*    /xsgc && do { $re_pat .= '.*';          redo };
    /\G \\(.) /xsgc && do { $re_pat .= quotemeta($1); redo };

    /\G \Z /xsgc or redo;
}
[download]

I suggest getting all the files once, and then grep'ing that list twice for your @cat1 and @cat2. I'm not completely sure what you're trying to match, but something like this?

my @files = $ftp->ls();
my @cat1  = grep { /h+_/ } @files;
my @cat2  = grep { /-zips/ } @files;
[download]