What are the simplest possible cards that could exist in your game? They should exist, because their complexity is super-low, and their gameplay value might be reasonable. Many of your cards should be nothing more than an icon. Most should be one sentence, and maybe an icon. Next, do cards that have two icons on them, or an icon and a sentence of text. Then, two sentences.
Look at every rule of your game, and make a card that breaks it in some way.
Not all of these cards will work, but you should try them.
Strategic depth should come from the interaction of many simple parts, not from the internal complexity of complex parts.
In addition, most of the deck should be composed of cards that do very straightforward things. They shouldn't be situational or unusual or flashy. They should be things you'd usually want, even if those things are boring on their own.
With fringe effects, that require specific circumstances, you can merge multiple of these fringe effects into one card.
When players draw a hand, the hand as a whole should be very easy to understand, useful, and maybe have one wacky or exciting card.
A deck is a deck, not a collection of cards. Most of its contents will be boring and functional.
To further simplify the deck, and make it more functional, I skew card decks heavily towards the aforementioned simple & straightforward cards. I put three to five of those cards in the deck, and only one or two of each of the exciting and strange ones.
I try to keep decks as small as possible, especially if the deck isn't the core of the game. You can easily make a deck much better, by halving its size, and keeping only the best half.
There's also another axis to consider here: predictability. I like to add strategy to games with cards, by giving the players some idea of what cards the opponent will play. If there are some common cards in the deck, a player can anticipate them, and strategise accordingly.
If the deck is full of large numbers of unique cards, players won't be able to strategise at all. The opponent's cards will just hit them out of the blue, and it will be feelbad.
Cards do a lot of good things for a game. They provide choices. They're a well of depth and replayability. They help tie turns together. If the cards are in players' hands, they're hidden information, and therefore be an easy way to avoid everything being calculable.
Most games should have a deck of cards.
Avoid "deck construction" (this is where players can modify or build their own deck, for later play.) Deck construction is not "just like a regular card game, but better". It's a massively-involved metagame on top of your game. It also implies a much greater content offering beyond the game itself, to enable the full spectrum of deck-construction possibilities. The vast majority of players don't want to interact this strongly with a game, or engage with such an involved metagame. If they do, they can just play Magic, not your amateur monstrosity. Deck construction should only exist in your game if it's central to the play — not as an add-on to a regular card game.