perhaps you can use biomart at ensembl: http://www.ensembl.org/biomart/martview/ or directly from biomart: http://www.biomart.org/

If you are able to code have a look at: using Bioperl to retrieve multiple sequences

Just change the query variable accordingly.

You can download the nucleic acid sequences in the GENBANK format using the following piece of code:Run through a for loop for many acc. nos. Then one by one parse the files to get the CDS.

use Bio::DB::GenBank;
use Bio::SeqIO;
use strict;

my $acc='NM_001003933';#example entry
my $format='genbank';
my $file_name=$acc.'.gbk';

my $seqout = new Bio::SeqIO( -file => ">$file_name", -format=>$format);
    my $getseq = new Bio::DB::GenBank;
    my $seq = $getseq->get_Seq_by_acc($acc);
    $seqout->write_seq($seq);

This piece of code can parse the GENBANK files to get the CDS in fasta format.However where CDS is not availble like in case of pseudogenes or rRNA etc this can crash. The scalar(@cds_features) will return zero then. Use it if you have non coding sequences

    use strict;
    use warnings;
    use Bio::SeqIO;
    my $gb_file="NM_001003933.gbk";
    my @cds_features = grep { $_->primary_tag eq 'CDS' } Bio::SeqIO->new(-file => $gb_file)->next_seq->get_SeqFeatures;

    my @tags=qw/protein_id product translation/;
    my ($feat_object)=@cds_features;
    my ($id,$product,$translation)=map{$feat_object->get_tag_values($_);}@tags;

print ">",$id," ",$product,"\n",$translation,"\n";