If there is one evidently best strategy the game is not balanced. Similarly, to make combat interesting, it must have meaningful decisions. Hopefully you can find ways to achieve that in this answer.
So, if the players are just doing the same thing without thinking, that is not good. Thus, our first stop is puzzle design. How can we make a battle into a puzzle?
Good puzzle design shows a path forward to the player, so they are not immediately stuck and can begin exploring how the puzzle works. But this evident path is not the solution. It leads to a road block, and the whole point is figuring out how to deal with that.
So we want a battle that changes kind of in the middle. The first tool we have for that are the skill costs in MP and TP. If MP depletes and TP accumulates during the battle, you will have the enemies doing a different set of attacks as the battle progresses. You can add to that sealing skill (e.g. via status effects or common events). And of course, you can put skills on turn times, or have the enemy use them when they have an status effect or low on HP.
And, of course the change needs to be significant. You want the late battle skills to be powerful. You, of course, test that this is working and the encounter is possible. At the expected level of the player characters.
That way the player has to adapt to the change. Or once the player knows it is coming, they need to have an strategy to deal with or to avoid it. Now the player is thinking.
In particular for a bosses, another part of puzzle design is letting the player know when their approach is right. For the RPG battle that is in good writing. Script some lines that clue the player on the enemy weaknesses and resistance. Perhaps the enemy says something during battle… So here is an idea: have a common event for each possible weakness, and call the common event from the relevant skills. Then on the common event you set a switch, and now you can check that switch on the battle events of the enemy troops and use that to trigger lines.
The problem with a puzzle, is that once you know the solution you are just going through the motions. Consequently that kind of design is fine for a throwaway boss, but not much else.
Our second stop will be strategy design. Good strategy design has three phases: Planning (and preparation), Practice (execution of the plan), Improvisation. Puzzles are focused on the practice phase. Let us broad the view:
- Planning, of course, happens in the player mind. But the games has features to assist it: we can see our inventory, we can see the stats of the party. And for preparation we might buy items, switch equipment (or go grind levels).
- Practice, happens in the battle proper. We came up with a plan and we execute it.
- Improvisation also happens during battle. It is when our plan went wrong, and we have to come up with a way to win on the fly.
Let us start on planning (and preparation) because there are interesting decision that (can) happen there. First and foremost: what equipment to use? But also what consumibles to buy and what party remembres to bring.
If the weapons and armor that you give the player are always straight up upgrades, there isn't much to decide. Instead build trade-off on the equipment. For example, in an area with a lot of enemies weak to magic, you want weapons with magic attack, but you make weapons with high magic attack physically weak so they don't work well in areas with a enemies that are weaker to physical attacks. And then you make areas with a log of enemies weak to magic attack, and a few strong enemies weak to physical attacks… And now the player needs two strategies, with only one set of equipment. Add elemental weaknesses and status resistances and you can add more variaty.
For practice, from an strategy point of view, as Phillipp was saying, give the characters multiple options of what to do on their turn. Since the battle is turn based, there is an action economy (each action has an opportunity cost). I want to go beyond that: what synergies can party members have?
But before we talk about synergies, we should talk about roles in combat. Please don't put each party member in a neat box. They should be able to fill at least a couple roles each. Otherwise the player just plays the character as their role and there is no interesting decision. Similarly, you build the party with redundancies. For example: have two characters that can heal, but with different trade-offs, one could heal a lot only one ally, and another heals a little but all of them, or uses regeneration instead of healing, or whatever. And remember that due to the action economy if they heal they can't do something else. And do that for each possible role in combat. So who is the player going to use to heal?
Alright, we have roles. Now I give you a tool for synergies: If you have states that increase element rate or state rate, you can make an attack or skill that imparts one of these states, making it easier for the skill of another character to work. And you can make that other remove the state. Have a few options of these synergies to use for each party combination (because, again, there must be a decision for the player), but interlocked in a way that if you are using one, you are not using others (by taking advantage of the action economy).
By the way, something very effective you can do for skill design is have skills cost HP. In RPG Maker MZ that means making using a common events to reduce the user HP, and to disable the skill when the character HP is too low.
Finally, improvisation. So you have this character that cast flammable, which makes enemies weak to fire, and this other character that casts fireball. But the enemy has deal significant damage and you need to heal… But the character that heals is the same that casts flammable. So you can't cast flammable this turn. Thus the fireball will not work. What now?
Consumible items will allow your characters to temporary fill a role they should not (you can make a healer out of healing potions as long as you have healing potions). So the consumibles are tools that the player can pull from when they need to improvise their way to victory.
Out third and final stop is boss design. It goes without saying, but just in cases: a boss should not be a random encounter.
A good boss (in particular in late game) should break the expectations of what enemies can do. So, use events. Here some ideas of what you can do: have the boss seal skills, steal items (events), summon other enemies (events), change its own weaknesses, remove (de)buff and status effects.
Here is an idea: give the boss an attack that inflicts an status effect that gives you restrictions, but also allows you to resists other of the boss attacks. Once the player figures it out, they have to decide if they want to cure the negative status effect, or keep it.
Anyway, breaking expectations for the sake of it does not feel fair. Thus, telegraph the attacks. How? Again, events. For example, have the boss use a "preparing big attack" skill on turn 6, and then use "big attack" on turn 7.
The telegraph should also work as a opportunity window. Make the boss weak on that turn, or even make the boss cancel the "big attack" if the player does a particular damage. They are also a good opportunity to heal for the player if they are low on health. Should the player heal and try to tank the "big attack" or should the player try to cancel it, knowing that means they be low on health?
And don't forget that the boss should be a challenge and a test. The player should not be able to beat the boss with regular attacks. Instead the player should have to use the tools they have, and use them well. Use phases to test to give different challenges to the player. Furthermore, you can design the boss phases on what the player must have acquired recently. If you don't have something more to challenge the player, don't put a long health bar on the boss (bullet sponges are not great bosses).
And of course, a good boss is significant for story reasons too. So don't waste the writing opportunity, not only in scripting lines, but also in skill and unique status names. Oh, and the presentation! The boss battle should be an spectacle.