Each turn should be a step in a larger plan, not a self-contained puzzle.
Just as it's important to tie game systems together with interlinking, you should also try to tie turns together, into a bigger, unsolvable game.
You want players intuiting a long-term plan, not just getting the most benefit in a series of self-contained turns.
Permanent benefits make the players plan long-term. This is another benefit of ramps.
Delayed effects of all kinds are also good. Delays might seem strange or bad, but they strongly tie a turn into the following turns.
In Radlands, you can play "event" cards. Each one has a "time" value, and will take that many turns before its effect occurs. Also, when you put a person onto the table, you can't use their ability that turn.
"Money" is almost always a good thing. It should probably be in your game. And, make sure to include some expensive things in your game. Buying something big is also a "big move". Just letting players save up, is a very easy way to tie the turns together.
Anything that lasts beyond the current turn is good.
Multi-stage actions are another way to tie the turns together. If it simply takes multiple turns to put all the pieces together, players must formulate a larger plan.
In Agricola, you just farm crops and animals, to produce food for your people. However, each of those two things is broken down thematically into many steps. To eat crops (efficiently), you need to get a grain token, plow a field, plant the grain in the field, wait for the harvest (whereupon you gain grain), get resources, buy an oven, and then bake the grain into bread. This very sequential gameplay means that you have to do the whole sequence, to gain any benefit. This makes Agricola a much more unforgiving, tense, and strategic game than most games in its category.
Hands of cards do all kinds of good things, and should be present in most games. They can be great for long-term strategy. Just make sure the cards don't come and go too quickly.
Race for the Galaxy is a great game, but I think it leans a bit too heavily on the tactical side. To play cards, you pay for them by discarding other cards. This too often means that you're discarding your entire hand. This breaks the connection between turns.
For most games, a steady drip of a new card per turn is good. It adds some uncertainty, and prevents concrete long-term planning. For games that revolve around the hand, turn that drip off, and force the player to play multiple turns without drawing a new card.
In Libertalia, players have a hand of nine cards. Each is a pirate ship crew member. Each player plays one of these cards face-down onto the ship, then they're turned over, and they interact, in amusing ways. Many also interact with the characters the player has previously played, which helps tie the turns together
even more.
Another key point here is that the players do not draw a replacement card. They play a total of six cards this way, and then
draw six new ones. This forces the players to plan out their hands, as
they'll have to play six of the cards, in some order. If players simply drew
a replacement card each time, they couldn't and wouldn't plan this way.