I've used CodonCode Aligner and Geneious previously and both seem to work well at this, however they are commercial apps and I'm not sure if the trial versions will work with any more than a few sequences (although Geneious seems to have quite a lot of free functionality).

You could use BioPerl's Bio::SimpleAlign to generate a consensus from an MSA, or BioPython's Bio.AlignIO module.

Perl:

use strict;
use warnings;
use Bio::SimpleAlign;

my $alignment = Bio::SimpleAlign->new('alignment.fas');
my $consensus = $aln->consensus_iupac();

Python:

from Bio import AlignIO
from Bio.Align import AlignInfo

alignment = AlignIO.read(open("alignment.fas"), "fasta")
summary_align = AlignInfo.SummaryInfo(alignment)
consensus = summary_align.dumb_consensus()

Try MEGA..http://www.megasoftware.net/