I take it you're talking about the standard unicode characters and not some odd thing from a specialized font, right? this should be fairly easy, either in applescript or by using textwrangler, I just want to make sure that I understand the parameters of the probllem. I'm wary of the fact that EPUB is misrepresenting the characters - that suggests to me that the problem may not be as straight-forward as you're presenting it; EPUB ought not to have a problem with standard unicode.
Here's some observations -
æ = U+00E6 = <c3 a6> (UTF-8) = <c3 a6> (MacRoman) = √¶ † = U+2020 = <e2 80 a0> (UTF-8) = <e2 80 a0> (MacRoman) = ‚Ä†
which likely mean that your source data is text in UTF-8 but destination (or viewer or intermediate converter) is interpreting the data as text in MacRoman.
Make sure you properly declare the encoding of XHTML or EPUB document as UTF-8.