Another approach is to remember the length of the longest match "so far" and the corresponding target.

use strict;
use warnings;

my $string = 'this is a nutrition assessment';
my @dic= (
 ['animal source food','aliment d’origine animale'],
 ['balanced diet','régime alimentaire équilibré'],
 ['food','aliment'],
 ['nutrition','nutrition'],
 ['nutrition assessment','évaluation de l’état nutritionnel']
);

my $longest = 0;
my $best_target;
foreach my $row (@dic){
    my ($TermDBSource, $TermDBTarget) = @$row;
    my $len;
    if ($string =~ m/($TermDBSource)/ and ($len = length($1)) > $longe
+st) {
            $longest = $len;
            $best_target = $TermDBTarget;
    }
}
print $best_target, "\n";



OUTPUT:

Θvaluation de lÆΘtat nutritionnel
[download]

UPDATE: Modified code to be more consistent with original post.