During the Second Swiss Estonian Workshop on Computational and Theoretical Linguistics one of the project groups trained a tagger for Estonian. It is based on Brill's Tagger. The tagset (developed at the workshop) consists of 63 tags (13 for punctuation). The tagger has been trained on a manually tagged/corrected corpus of estonian fiction consisting of about 7'000 words.
The tagger is in an initial experimental state and can be tested here.