There seem to exist a lot of different options with very dfferent complexities, paired with different word-formats.

If it's a plain old word 2000-2003 file and you already know what your tables look like and you need only some data from within some cells, you could do simply a:

$> abiword --to=rtf myworddocument.doc
[download]

$> perl extract-table-cells.pl myworddocument.rtf
[download]

in the latter (extract-table-cells.pl), you would simply search for:

[pseudo] 
 ...
 # table content part already extracted to $tablecontent
 @cells = $tablecontent =~ /} ([^}]*) }\\cell{/xgs;
 ...
[download]

which might give you the cells in @cells.

But it depends on your problem. Of what scale and purpose is your attempt?

Regards

mwa