As Kanji says, HTML::Parser is the most common choice for parsing HTML.

Another option that is good in some situations is HTML::TokeParser which uses a different (and some say easier) approach.

HTML::TableExtract is also handy for the more specific task of grabbing text from HTML tables.