Here's a single pass through your  @big array and should be fairly fast:

c:\@Work\Perl\monks>perl -wMstrict -le
"use Data::Dumper qw(Dumper);
 ;;
 my @big = ('d:\1\1\1.txt', 'd:\1\1\2.txt', 'd:\1\1\3.txt', 'd:\1\1\4.
+txt');
 ;;
 my @small = ('\1\2.txt', '\1\3.txt');
 my ($has_some_small) =
   map qr{ $_ }xms,
   join q{ | },
   map quotemeta,
   @small
   ;
 print $has_some_small;
 ;;
 my @inte = grep m{ $has_some_small \z }xms, @big;
 print Dumper \@inte;
 print qq{'$_'} for @inte;
"
(?^msx: \\1\\2\.txt | \\1\\3\.txt )
$VAR1 = [
          'd:\\1\\1\\2.txt',
          'd:\\1\\1\\3.txt'
        ];

'd:\1\1\2.txt'
'd:\1\1\3.txt'
[download]

As a beginner, always  use warnings; and  use strict; at the start of your scripts. See warnings and strict. Data::Dumper is core, so you should have it; it's a great aid to debugging. Please see perlre, perlretut, and perlrequick for regex info.