$to_email =~ s/.+To:\s(.+?)xxx.*/\1/;
[download]

In addition to the non-greediness change suggested by jethro, you'd also need a trailing .* to match the remainder of the line, because only what matched will be subsituted.

Also, you'd need to take care that the 'xxx', which you use to locate the end of the mail address, will not itself be part of the mail address.  (Not sure of how much relevance that potential problem is here, so as a quick fix, I simply changed .*? into .+? — the .+ consumes one of the x'ses, so the first occurrence of 'xxx' will no longer be a candidate for the terminating 'xxx'...)

This works! Thank you and everyone that helps...

ps: The xxx is just a placeholder. The mail adress itself will be different. It is my fault to replace the line endings with xxx's and using the same for the e-mail :)

$to_email =~ s/.+To:\s(.*?)xxx/\1/;
[download]

will do what you want. The * is greedy which means it will take as much as it can. The '?' changes the * to a non-greedy version

Thanks. But the result changes slightly

xxx@yyy.netSubject: Size atanan xxx numarali ticket hakkindaxxxContent-Type: text/plain; charset=ISO-8859-9xxxContent-Transfer-Encoding: 7bit

(the xxx near the e-mail address disappeared)

TIMTOWTDI (unnecessarily verbose, but perhaps easy to follow and illustrating some possibilities with lookaheads):

#! /usr/bin/perl
use strict;
use warnings;

my $text = 'From: aaa@bb.netxxxMessage-Id: xxx<200902131530.n1DFUZ5l00
+1303@aaa.bb.net>xxxTo: xxx@yyy.netxxxSubject: Size atanan xxx numaral
+i ticket hakkindaxxxContent-Type: text/plain; charset=ISO-8859-9xxxCo
+ntent-Transfer-Encoding: 7bit';

if ( $text =~ /(?=To:\s)(.+)(?=[^x]{3})/ ) {
    $text =~ /.+To:\s(.*?)(?=xxxSubject)/;
    my $to_email = $1;
    print $to_email . "\n";
} else {
    print "no match\n";
}
[download]

Output:

as specified.

Your OP would have been easier to parse had you included data and code inside <c>...</c> tags as done above.

Also, it is not clear if the text you give in your post is multi-line, i.e., if it has embedded newlines, or if it is a single string (which seems most likely). In the former case, use a /s regex modifier (see perlre section Modifiers) which allows the  . (dot) metacharacter to match newlines.