Here is the crucial question: are your dungeons good?
Because if the answer to that is, "Yes.", then I can teach you how to do every single other sort of adventure simply by reapplying what you know about dungeons. Really, it's a trivial excercise, and the better your dungeons are the easier it will be to create something else.
If on the other hand, you are doing dungeons to disguise the fact that you can't do anything but random encounters, you've got a deeper problem which is - you have a basic problem thinking up ideas. If this later part is your problem, then the best course of action will be steal ideas and make them your own. I would suggest starting with published modules and spending some time making them your own. Do some self-analysis.
I'll procede for the moment with the assumption that you make good dungeons, and I'll briefly outline how you go from dungeon to anything else.
The basic thing to remember is that every encounter of every sort is a room, and every choice is a corridor. People gravitate to dungeons primarily because the number of choices they present the players is small, so its easier to plan out the adventure ahead of time. With every other sort of adventure, the trick is managing the 'corridors' so that the 'map' of the adventure is simple enough to plan for.
So, you want to do a wilderness adventure, the main points of interest are the rooms, and they are logically linked by corridors according to the sort of graph paper you use for the map. That is, on a square grid, every room is linked by four virtual corridors going North, South, East, and West. If you want to get fancy, then you have NE, NW, SW, and SE, but the judicious use of terrain can be used to effectively block these options if you need to. A simple wilderness adventure might be 'Desert Island Survival'. Your points of interest might be:
a) Rocky area where shelter and fresh water can be obtained.
b) Mountainous area where terrible monster lurks and treasure can be obtained.
c) Swampy area where there is risk of disease and minor monsters.
d) Lagoon where fish can be obtained.
e) Grove where fruits and coconut can be collected by day.
f) Bamboo grove gaurded by minor monster. Raft and other tools can be built here.
g) Rocky area where flint can be obtained for starting fires and making simple stone tools.
All that remains is to connect the areas in a way that is both logical and which develops the story in an interesting way. The story is here primarily one of 'Man against Nature', and you want to arrange the island in such a way that the protagonist must make hazardous journeys across the island to obtain the tools needed first to survive, and ultimately to overcome the terrible monster and escape from the island. This suggests corridors such that, for example, the adventurers must traverse back and forth between the places where shelter can be obtained and those where food can be obtained. Eventually this leads to conflict with the terrible monster, and the need for weapons sufficient to defeat it. This eventually leads to the monsters destruction, the reward of some treasure, and hopefully some eventual escape via raft to some nearby island where some further adventure awaits (for example, the natives know where foreign merchants can be found, but first the players must defeat the nasty witch doctor and his zombie minions).
The island is a dungeon. It has rooms - big rooms, but rooms nonetheless and corridors between them. You just don't describe it that way. You describe it as an island. Behind the screen you are tracking the PC's movement in the exact same way you would in a dungeon.
Or suppose the adventure is, "Survive the assault of orc mauraders." and the setting is a one room farmhouse. This is a very simple event based dungeon. It's worth looking at because it is to event based adventures what a dungeon is that has a single line of rooms connected by a single door. It's linear dungeon. It has a linear map. In the case of the event based dungeon, all the rooms are identical, just the content changes. The players don't have any real choices here. After defeating the monsters in room #1, they enter room #2, and so on. The interesting choices are how do they handle the tactical problems presented using their own resources and the tools you've put in the farmhouse (or its environs).
Of course, just as you can make the map of a dungeon have any degree of complexity, you can make an event driven map be as complex as you like as well. You could have for example, six separate locations each of which has a series of events like the farm house, and going back and forth between the locations determines which rooms the adventurers see. Further, you can put forks in the dungeon, so that depending on choices they make in a particular event, the rooms available to them change. However, if you draw it out, what you end up with is a map that differs from a dungeon only in that it has alot of one way corridors.
If you keep in mind that everything is a dungeon, then you'll be ok. Apply what you know about making dungeons to making things that don't look like dungeons (but are!) interesting, and you'll be fine.
), or a small abandoned village (a couple dozen buildings, half of which are crumbling, leaving the other half for little things to explore making it an outdoor dungeon confined to the village perimeter)...
