The only weapon that comes to mind at the moment would be some antimatter-based weapon, but that weapon would have an enormously dangerous drawback that would make it impractical, which is the fact that if there were some mishap that caused the antimatter to come into contact with matter, that the weapon would then go off. (Nuclear weapons, on the other hand, are typically designed so that it is extremely difficult for them to go off by accident.)
Also, the fact that antimatter is extremely difficult and costly to manufacture.
Maybe some weapon that harnesses the Sun’s energy into some form of weaponized use?
Collecting large asteroids and stationing them in orbit around the earth, then dropping them onto cities would be as devastating as you want. It isn’t feasible, but it’s more feasible than antimatter.
The classic scifi example is asteroids or some other sort of kinetic impactor from space. Launching kinetic impactors on an ICBM would be silly though, since only a small fraction of the chemical energy in the rocket would be delivered to the target. The most “practical” option would involve nudging a near-Earth asteroid. Done properly, and the energy it takes to nudge the asteroid will be far less than the energy released on impact.
The best weapon yet would be to deliver toxins in small quantities that would wipe out whole cities. Achievable with current technology.
Depends on what you mean by ‘powerful’. Biological weapons, simpler poisons, and nano-tech could kill everyone on earth without all the physical destruction of nuclear weapons or dropping asteroids on the earth. Conceptually the ozone layer could be destroyed chemically and perhaps upset the ecology so greatly that human life is barely feasible.
How about a bigger nuke? There’s no practical upper bound to how big you can make a fusion bomb. Or rather, there is eventually an upper bound, but it’s far greater than anything we’ve ever actually made.
There’s not much point, though. With the bombs we already have, your target area is completely flattened, and the size of your target area is limited by the distance to the horizon, not by the energy of the device.
Actually, from what I’ve read, the best use for a small amount of antimatter wouldn’t be as a bomb, per-se. It would be most useful as a boosting agent for existing fission weapons, or to produce very low fallout fusion weapons, as the matter/antimatter reaction would give off a lot of energy and neutrons if you were annihilating antiprotons with uranium or plutonium, and would fission those elements much more efficiently, allowing for smaller amounts of fissionable material.
Also, the nature of the annihilation event means that the energy given off could be used for a standard fusion secondary without the need for a fission primary (where the transuranic fallout elements come from).
But **Willcross **is right; dropping stuff from orbit would probably be the cheapest/easiest way to do something as powerful or more so than nuclear weapons. Certainly cleaner anyway. Something as light as the Space Shuttle could lift (25 tons) would have an equivalent yield of about 25-30 tons of TNT, and there aren’t any real countermeasures against such a thing.
A real asteroid would be correspondingly more powerful as its mass goes up.
25-30 tons of TNT is far weaker than the Hiroshima bomb, and it’s extremely expensive to have spacecraft lift a 25-ton object into orbit. I do not see how that is more cost-effective than the ICBMs that already exist today, where a Minuteman can deliver several dozen Hiroshimas.
Actually there’s an even simpler practical option, which the US has toyed with as a concept. They called it “Rods from God”.
The idea was you would load up a satellite with some rods, which are telephone-pole sized hunks of tungsten, then drop 'em as needed. The ones envisioned would have been somewhat small and used as bunker busters or very low-yield nukes ; but there’s no practical reason not to scale the concept up some, and then some more. After all, the only bottleneck to that method is how much a rocket can carry into orbit ; and you could even have multiple payloads sent up and reassembled later.
Well the other bottleneck is sanity, but 
This was my first thought. They even depicted them in the otherwise abominable GI Joe: Retaliation. The movie depicts one being used to create an earthquake that levels London, IIRC. I don’t know that they would necessarily create a seismic event as opposed to a simple explosion, but the fact remains that dropping something from orbit is probably the best way to increase kinetic projectiles to megaton yields.
The Tsar Bomba was 58 megatons. How big would these rods have to be to exceed that?
How much does that space station weigh?
That is an important point. If the OP question said “more powerful”, but specific designs for nuclear weapons have been studied to at least 50,000 megatons. Those could actually be built and the OP question is what alternative weapon which could be built is more powerful? I cannot think of anything.
Although there is no theoretical upper limit to fusion bombs, there is a gradual practical limit based on yield-to-weight ratio. This upper limit is 6 megatons per metric ton of mass: Nuclear weapon yield - Wikipedia
So a 50,000 megaton device would weigh 8,333 metric tons and would have to be delivered by ship. The largest bomb a Saturn V could lift would be about 816 megatons (136 metric ton LEO payload * 6 megatons per metric ton mass).
The largest bomb which could be airlifted would be 1,518 megatons by an AN-225: Antonov An-225 Mriya - Wikipedia
Of course due to the inverse cube law, it is extremely inefficient to use a single large bomb – whether conventional, nuclear or some hypothetical antimatter device. For a given payload mass it is much more efficient to use multiple smaller devices. This can be seen by experimenting with various yields on this simulator:
If using a small amount of antimatter as a potentiator for nuclear weapons is allowed, then extraordinarily powerful “weapons” can include, as we well know, box cutters.
I’m thinking setting off somehow an earthquake or tsunami; I think blowing up a dam (SD has explored this theory with the Aswan Dam) would be perhaps even more destructive.
Causing a criticality or other form of radioactive realize at a nuke plant would be less extensive but more long lasting. Of course sometimes your enemies don’t have nuke plants around.
How hard would it be to inflate a giant lens in orbit? A small magnifying glass can be used to torture insects; a large lens focused on a Daesh stronghold might cause considerable discomfort.
Megadeth: Breaking the Aswan Dam
Dropping things from space is only practical if you got the things from space to begin with. If you’re just going to launch them yourself with a rocket, then you might as well use the rocket itself as your bomb: It’s at least the same energy, and in practice a whole lot more (because rockets are really inefficient). But of course, rockets (current ones, at least) get all of their energy from chemical reactions, which fall far short of nuclear reactions.
Well yeah, but you can use multiple rockets over time to bring more mass into space that can be carried on a singular rocket. It’s very much not cost-efficient, but military stuff generally isn’t anyway. The F-35 is my cite :D.
The ISS has a mass of 419,455 kg (924,740 lb)-close to a half million tons. Is that large enough?
If you define a ton as two pounds, sure.