As an example, let's say you define your promoter as the 2kb upstream of your gene and the you have a bed file with the chrom, txStart, and txEnd, name, num_exons, and strand for each gene you are interested in. Something like the following:

head -n4 genes.bed
chr1    134212701    134230065    Nuak2    8    +
chr1    134212701    134230065    Nuak2    7    +
chr1    33510655    33726603    Prim2,    14    -
chr1    25124320    25886552    Bai3,    31    -

bedtools flank -i genes.bed -g mm9.chromsizes -l 2000 -r 0 -s > genes.2kb.promoters.bed

This will give you the upstream regions based on strand as follows:

chr1    134210701    134212701    Nuak2    8    +
chr1    134210701    134212701    Nuak2    7    +
chr1    33726603    33728603    Prim2,    14    -
chr1    25886552    25888552    Bai3,    31    -
`

You can now use this BED file to extract the sequence (based on strand) from the mm9 genome.

bedtools getfasta -fi mm9.fa -bed genes.2kb.promoters.bed -fo genes.2kb.promoters.bed.fa

NOTE: The "mm9.chromsizes" file is a tab delimited file where each line has a chrom name and a chrom length. See the bedtools manual for examples. mm9.fa is meant to represent the name of the mouse reference genome in fasta format.