Random mutation. If it reproduces successfully, those genes pass on. If successful enough, it could become the dominant variety. If not successful, that variety will be a small population or die out. That's all there is to it.

In addition to what other people said, I'll focus on the delicious fruits.

A lot of wild fruits are awful when raw. Crab apples are a good example - small, tough, excessively tart. But then you get humans picking the least awful of those fruits, and spreading their seeds (sometimes without a thought, sometimes on purpose), you're effectively selecting the best-tasting ones. And across multiple generations, the fruit goes from barely edible to passable to okay-ish to good-tasting.

In other words, plenty tasty fruits out there are not the result of trees "trying" to propagate themselves better. They're the result of weird monkeys doing artificial selection across millenniums. Or even through the centuries, as this Renaissance painting shows:

Check the bottom right - those are watermelons. Granted, watermelons aren't from trees, but the reasoning is the same - in just three centuries or so watermelons went from "90% styrofoam with some good bits" to big balls of juice. Why? Human intervention.

They don't know. They exist because their traits allowed them to survive better than other traits. If an evolution fails, they cease to exist. What's left over is what you get to see

There is no knowing or feedback.

Mutations are random and there are many more failed mutations than successful ones.

You only see the successful species for that very reason, the failed ones have not been able to compete or pass their genes on at an appropriate rate.

It's the same as any other evolutionary process. Those who survive and reproduce the most carry their genes to the next generation; those who are less well adapted don't reproduce as much and over the generations their genes get removed.

Say you have a bush with berries. Some bushes make sweeter berries than others. Let's say those with the sweetest berries will attract more birds and animals who will eat the berries and scatter the seeds further away. The sweeter bushes are getting more lotto tickets for the next generation, the sour ones not so much. Repeat over time and eventually the sour ones will be gone.

The same can be applied to any other trait. It depends on the pressure of the environment they live in.

It's not that complex.

A mutation happens. It can result in either a disadvantage that prevents the plant from surviving to reproduce or an advantage that helps the plant survive to reproduce.

There's no "knowing" involved. There is no intent on the part of the plant. It's just random. You're seeing the descendants of the plants that got an advantageous mutation.

What you're not seeing in your yard are the billions of different plants that never survived because their mutation was a disadvantage.

"What survives survives, what doesn't doesn't."

As others have said, evolution doesn't really work that way: an individual organism doesn't make changes to suit its environment. Offspring aren't carbon copies of their parents, there's variation of pretty much every attribute. Most attributes don't really impact survivability, but sometimes they do. If certain things tend to result in an organism surviving to reproduce better than other things, those traits get reinforced and are more likely to be passed on. Over lots and lots of generations, you can end up with pretty much every organism in that species having that trait, and that trait getting more and more pronounced.

It doesn't. The fruit that gets eaten gets to make more via the spread seeds. The plant that's bitter doesn't get eaten before it can go to seed. It's not more complicated than that.

How do they know it worked/failed.

A lot of answers here, but I think what’s still missing is the most intuitive. The plant that sent the seed doesn’t know, rather its offspring does - by simply surviving. The species as a whole “knows” by whatever works.

How do they receive feedback about their evolutionary experiments? How do they know it worked/failed.

They're not performing experiments, mutations are random and they either help the plant survive/reproduce, make it harder to do so, or have no appreciable effect. If it helps the plant, more of them will survive and reproduce, meaning more of the plant with that mutation will be more common. If it hurts, then it will be less common.

I'm with you. My wife has REALLY gotten into plants in the last year or so, and it amazes me how "smart" they are.

Obviously, everyone here is right. There is no intelligence, just genetics, but watching my morning glory wall climb the rope net to the roof of the house just blows my mind.

It's crazy to think about all the trial and error over the centuries that it took for a simple flower to develop little sensor hairs that explore its surroundings and wrap itself around anything in range.

If, over a long time period, that tree's particular characteristics lead to the species surviving, then it worked.

If it doesn't, then the species dies out.

There's no feedback, other than whether or not the species continues to exist.

Taste and poison are easy to account for, it's way weirder to see creatures that die immediately after sex.

Genetics.

Your senses do not feed back into evolution. It's all random mutations that happen to make it slightly better at surviving. Generational survival is the pass/fail signal.

For many things that seem to require a large single leap in progress it turns out that there is a clear story of gradually developing it over a long time.

You might like Richard Dawkins' book The Greatest Show On Earth

How do you make eggs or sperm?