Creating a hash is a good way. You can sort the result for each id, so you do not have to search for the same ids.

my %by_id;
push @{ $by_id{ $_->{id} } }, $_->{score} for @data;

# Sort the scores for each id.
$_ = [ sort { $b <=> $a } @$_ ] for values %by_id;

# Sort the ids by the highest score.
for my $id (sort { $by_id{$b}[0] <=> $by_id{$a}[0] } keys %by_id) {
    print "ID: $id\tScore: $_\n" for @{ $by_id{$id} };
}
[download]