I generally find that it takes less willpower to execute a plan that I've already made. I set aside a little time every morning, and a longer period every Sunday, to be effortfully strategic and come up with some specific next actions that I can mindlessly execute for the rest of the day/week. I think this is more or less standard GTD (although I've been iterating on my personal system for long enough that I can't really remember exactly what David Allen describes).
I agree that separating 'planning' and 'doing' like this works especially well for doing aversive things. Your 'planning self' doesn't have to worry about actually doing anything, and your 'doing self' just has to trust your planning self.
Here's an example of a mental manoeuvre I accidentally found, and thought might be generally useful (typical caveats apply).
I've had a manageable-but-important Problem for a few months now (financial in kind, details neither relevant nor interesting), of moderate complexity and relatively minor importance unless I leave it unsolved just a little longer.
Unfortunately, this seems to be the precise combination of things that triggers one of my ugh fields, which manifests subjectively as a fuzzy blank inability to maintain focus. Several times last week, it occurred to me that I should really Solve The Problem, but I wasn't able to get myself to spend any time thinking about it. Like, at all.
On Saturday, the Problem found itself top of mind once again. How irritating that I couldn't solve the Problem because it was the weekend, and when it wasn't the weekend, maybe Tuesday when work wasn't busy and the Bureau was open, I should really email Dr. Somebody and call Mrs. Administrator for the ...
*blink*
I had a solution, and a plan. What the what?
My working theory is that when there's no chance of actually Doing Something, this particular ugh field deactivates.
To me, this suggests a strategy (of uncertain generalizability): when an ugh field is preventing thought about something important, find a time when action is impossible and use it to generate a plan.
I would feel better about this advice if it had a deep theoretical backer. Anybody?