As a means of poetry, I’ve so far only used syllable count.[1. Exemplified in my take on Shelley’s poem Ozymandias and the LCC4 relay (PDF)] What about rhyming word stems, though? For example, karon ‘water, sea’ and beson ‘ship’ rhyme – /rɔn/ : /sɔn/. Could they still be considered to rhyme even if arbitrary suffixes were stacked on them (or not), e.g.:

[gloss]Silvu besonyeley,
ˈsɪl.vu ˌbe.sɔn.je.ˈlɛɪ
see-IMP ship-PL-P.INAN
See {(the) ships,}[/gloss]

[gloss]Yam sarateng karon
jam ˌsa.ra.ˈtɛŋ ˈka.rɔn
DATT go-3P.INAN sea.T
To₁ {they go} {the sea₁}[/gloss]

[gloss]Mang’ avan nongonya.
maŋ ˈa.van nɔ.ˈŋɔn.ja
MOTION bottom river-LOC
Down   {(a/the) river.}[/gloss]

Or would that be too far-fetched? After all, in the case of beson and nongon, word stress shifts around wildly due to the added suffixes, which lessens the similarity in sound even further.[2. Note that all lines contain six syllables at least!]

I’ve so far avoided rhyming with suffixes because that wouldn’t really be too much of a challenge in terms of artificiality – it would be like using the same word twice to force a rhyme in English. On the other hand, it’s not like this wasn’t done in Latin (though post-Classical in that case), which prominently features suffixes as well. Though in the case of “O Fortuna,” the last syllable of a word stem is also taken into account, plus inflectional suffixes, creating a polysyllabic rhyme. Doing it this way would mean, though, that you’d have to make sure the rhyming words are inflected for the same grammatical categories, which in itself might be an interesting challenge as well.

On a completely unrelated side note, look what Miekko has been doing for the past three weeks: Miniature Conlangs.