One part of this is understanding how much you LIKED things. How do you compare your previous experiences in relevant reference classes to each other, and to your priors (based on others' reported experiences)? Your model of how to map outside views to actual predictions is differential, not standalone.
For a WHOLE LOT of such things, don't worry too much about prediction, instead just try it and focus on evaluation. How much are you enjoying this moment? How much did you enjoy whatever you did yesterday? What specific parts would you like to repeat, or change?
One of the things that I find helpful is to see how similar people who like X are to myself. I'm currently considering a career change and I can identify three important parts of this career. In one I'm very similar to people I know in this career, in another I find it hard to say how similar I am and in a third I'm quite dissimilar. And because of this I see a fairly big risk (I have high uncertainty about) me enjoying this career.
I'm willing to bet you've already heard of it, but it not you may want to read a bit about Affective forecasting. The gist is that we (humans) are typically bad at predicting what we'll like. https://en.wikipedia.org/wiki/Affective_forecasting#Improving_forecasts
They seem to hype mindfulness in that section.
Effort justification also comes to mind: https://en.wikipedia.org/wiki/Effort_justification
My take away from both is that context of how the decision to do X was made matters a lot... maybe more than X itself.
I might also suggest there's something like a "static fallacy" here (I made that up just now, but hear me out). You would expect your tastes to change and develop overtime as you're exposed to new things. In some sense X would be the same, but you or your relationship with X could change (e.g. like an acquired taste). There's also habituation to think about.
I'll even go a bit further and say there's a kind of joy in discovering that you dislike things. I think of RedLetterMedia's Best of the Worst series where the point is to enjoy bad movies. That probably extends to other things--have you ever asked a friend to try something that you found revolting and it was kind of fun to share the disgusting experience together?
I think what most people want to avoid are high investment buyer's remorse type decisions where you spend a lot of time or money on something and discover that it sucks.
Maybe the best thing to do there is a lot of testing the water. If you can, try small version of whatever it is before you overcommit.
Suppose you've never tried/done/seen/tasted X before. You want to predict how well you would rate/enjoy X. (X could be as small as a restaurant, or as big as a career choice.)
My knee-jerk way to approach this would be "(read about and) imagine doing X, see how much I like the thing I'm picturing, and form my prediction based on that." This approach seems to be subject to availability bias. Furthermore, this approach seems to take the inside view on prediction. For these reasons, I'm concerned that my knee-jerk approach will produce bad predictions.
On the other hand, how much you'll like something seems to depend more on 'inside' factors than whether you'll finish a project by a certain date. Does it make sense to throw out introspection entirely, and predict how much you'll like X based only on surveys/star ratings of X?