What is your opinion of the Deutsch-Wallace claimed solution to the probability problems in MWI?

Now we're getting into the philosophy of QM, which is not my strength. However, I have to say that their solution doesn't appeal to that part of me that judges theories elegant or not. Decision theory is a very high-level phenomenon; to try to reason from that back to the near-fundamental level of quantum mechanics - well, it just doesn't feel right. I think the connection ought to be the other way. Of course this is a very subjective sort of argument; take it for what it's worth.

Also are you satisfied with decoherence as means to get preferred basis?

I'm not really familiar enough with this argument to comment; sorry!

Lastly: do you see any problems with extending MWI to QFT (relativity issues) ?

Nu, QM and QFT alike are not yet reconciled with general relativity; but as for special relativity, QFT is generally constructed to incorporate it from the ground up, unlike QM which starts with the nonrelativistic Schrodinger equation and only introduces Dirac at a later stage. So if there's a relativity problem it applies equally to QM. Apart from that, it's all operators in the end; QFT just generalises to the case where the number of particles is not conserved.