Then allow ships to fly anywhere but if not within supply base range they get -50%penalty to damage (or tactics or whatever).
This is what I call an artificial rule. I generally try to refrain from using these arbitrary rules. These are rules that, while they allow to fill in an objective, don't fit organically in the design. In other words, the player wouldn't guess this outcome unless he was told about it through in-game messaging. As a counter-example, if the fuel/fuel tanks was a modifier to ship armor and damage, players would understand indirectly that, the further they get from home, the more vulnerable their ships would become. There wouldn't be a specific distance or specific value that would kick in and could be easily forgotten about and come back to bit the player in the ass: everything would be clear from the start and the player would understand that its part of the game economy and that he'd have to gauge his moves accordingly.
Now, I won't actually implement this system as its silly (I don't see how it could make sense to the player anyway) but it would be a more organic method to implement such a system.
Then don't allow/discourage taking a lot of fuel It's your game and your rules, you don't like something, don't allow it
Interesting you should mention that. Currently, fuel consumed per unit of time is based on ship mass, and ship mass encompasses fuel tanks. Thus, the more fuel you have, the further or faster you can go, but insodoing, the more fuel it costs. In the end, there is an advantage in bringing more fuel, but a portion of this advantage translates into burning a lot more fuel resources. From an economic macro standpoint, its a bad move, but the tactical/strategic advantage of grabbing a few planets is too dire a consequence.
Besides, I'm really interested in adding a layer of strategy to the game, and a defensive aspect would be necessary.
For example, a lot of medieval games use 'walls' and a lot don't. I generally like those that do because they allow you to craft a part of the landscape to suit your need. They give the player an opportunity to be creative about what they'll do (create a trap? give an incentive for the player to more through a specific choke? slow them down exactly where you need them to be?, etc).
Playing with ships is fun, but I'd really like to throw them another bone, a game ingredient/toy the players can use to mess with one another.
What if sensors/communications were a key part of the game? If your sensors are disrupted then you lose detailed information on what is going on in a region. Life on the planet would still continue, and they would carry on as they were, but if an enemy fleet with suitable jamming equipment got in range they could black out part of your map and allow a fleet to move without you knowing exact details.
Its actually part of the plan. I've got one of the races using a radar jammer sort of ship. In additional, planetary sensors are often found as satellites, and if broken, will be reduced in efficiency, rendering the player 'blind' to certain areas of his empire. That said, this is an offensive strategy, not a defensive one. I like the strategic layer it adds (there are specific targets you want to disrupt in order to come in blindly, or you can feign at attack by eliminating a few satellites and attack the other front instead), but it does not provide a means to defend oneself from the enemy though.
Combine that with the long duration battles to actually conquer planets, and it gives a viable defensive measures with attack counters:
I'm affraid I'm missing the part where there's an inclusion of an efficient defensive strategy here. Wouldn't the player play cat and mouse with his opponent, not knowing where to hunt them down?
Of course, relying on building a fleet of jammers means you give up resources that could have gone towards your attacking fleet, which puts you at risk of being clobbered that much easier if you run into the opposing battle fleet.
So essentially, you are taxing the attacker with the need to build radar jammers which are inefficient combat ships. That's also a side objective I'm currently pursuing. I've also included the need to resupply (ordnance, missiles, fuel) in such a way where the defender is always at an advantage (he is much closer to his bases). That said, it mostly means attrition of the attacking forces, but assuming he's waited for a critical mass, there still wouldn't be an efficient way for the player to have prepared for the defense. If you don't know where the enemy will strike, or if he chose to run by your forces and go straight for the heart of the empire, there will be little you'll be able to leave behind to counter this, and that's what I'm really after (the thing you can leave behind around important clusters when your fleet needs to be someplace else).
Area control and surprises. What if fleets weren't always easy to see, especially for attackers? Set up sensor girds to better detect vessels entering your region of space, so the defender has the advantage of knowledge.
Yes, there are recon vessels/scouts (even some cloaked) already planned. When I mean the enemy fleet runs past your fleet for a planet, I also mean they know what they are attacking, which inherently means they've scouted it before or are doing it currently. First rule of engagement dictates you need to have recon first.
Use some kind of webbing mechanic to allow a smaller fleet to delay a larger one,
How would this work exactly? This is much closer to the mechanic I'm trying to define, but it still sounds vague. I'm trying to get a concrete mechanic in place that does precisely that.
It could be building specific ships, as you seem to imply, whose purpose would be to slow down the enemy (VGA Planets uses minelaying ships in this manner) or it could be something else altogether. Also, slowing might not be sufficient. The threat of perhaps eliminating a few of your enemies is also important as it could really shift the battle in the defender's favor, forcing a retreat, or allowing the defender to trap his opponent and take a big win. There's little in life that's as satisfying as designing a trap (thinking about what the opponent will do, etc.) and seeing it succeed. It requires an investment of some kind (minerals, money, time, etc) and may yield favorable results (or not).
So in theory, what you're proposing works, I just need something more practical.

Find content
Not Telling