If you are familiar with how-to-use Perl module, this will be pretty easy. Basically you need to write a Perl script using the module and define your annotation file, background list, and the parameters. Various example scripts are provided with the GO::TermFinder package in the /examples directory. You need to execute them after modifying the path and include your own local files (perl analyze.pl < parameters >) to get the result files.

Example scripts that describe how to use GO::TermFinder are available here

I would also recommend you to take a look at the report generator example page.

I would partially agree with Lars suggestion, there are more easy-to-use tools out there, but the pro of GO::TermFinder it is still one of the most flexible GO enrichment calculation program in dealing with gene lists and defining background. It can be easily integrated to pipe-lines via Perl scripts to calculate enrichment for multiple lists. If the analysis is too specific for a well defined background, I would recommend a program that can be access a module or a BioConductor package which allows scripting and interaction to change the parameters.

My answer would be a question: why do you necessarily need to use the Perl module GO::TermFinder?

If your goal is simply to find enriched GO terms among a set of genes, numerous web tools exist that allow you to do so without writing any code. The official Gene Ontology web site provides a long list of such tools.

GO::TermFinder can show the enriched nodes in the original directed acyclic graph rather than destroy the structure, which is superior to other tools.