Just keep writing and deleting until you're happy with an idea, it's one of those things where there's no set method. Theory will probably help, but when it's all said and done you just sort of have to keep churning out piles of shit and eventually you'll come across an idea that's actually half decent. At least that's what I do, I've spent days just trying to come up with some half decent chords for some of my stuff.

I'm not an expert, but I've been having decent success recently. Keep in mind most of this is kind of setting up for huge supersaw chords, although it will also sound good with pads/piano/other things. I usually work in A minor so that it is easy to stay in key.

Note: You will need basic music theory knowledge - scales, chords, triads, inversions. Read up on these if needed.

What I've been doing is first sounding out a basic 4-chord progression that I like, just basic triads. Then once I have that 4-chord progression, I add a root note to each chord an octave below, so now each chord is four notes and it has some bass. From there I space out/invert the notes in the chord so that they are more spread out. (It helps to already have a good supersaw patch when you are spacing the notes out, you will be able to hear when it is "right").

Now with a wider, fuller chord sound due to the voicing changes, I figure out a rhythm for the chords that sounds good. I usually just loop the chords over a kick+clap and chop them up into something that sounds good.

Once that is set, you can play with the voicing of each layer of the supersaw. I often like to add another set of voices by adding a higher note on top of the chords that repeats for most of the progression, maybe changing up or down at the end.

I will also just take individual notes in the chords, try moving them up and down, and see how it sounds, and really just go trial-and-error following my ears to make small changes to the chord voicing. Also try adding notes in various spots and see how that sounds. Whatever variations you introduce, you will probably want to repeat them throughout the progression (so maybe chord 1 and 3 are varied in one way, chord 2 and 4 are varied in another similar way, etc), otherwise it may not sound "right".

It's helpful for me to think of each note in the chord as voicing its own melody within the supersaw, and once you can separate each note in your head, you can really make some really cool chord progressions without having to understand (in music theory terms) what you are doing.

Finally, once the notes are all set, try adjusting the velocity of each layer of notes. I usually reduce the velocity of 3rd/7th notes to emphasize the root/5th notes, and also will soften the highest notes if I've added another layer of notes on top of the chord progression.

I went through a phase of working out chord progressions from pop songs. I can play keyboards at a basic level, and I've done a tiny bit of music theory, but at the start I couldn't link that experience to understanding how a song actually works: I couldn't identify chords and keys by ear.

The way I did it went something like this:

First, try to work out the root note of each chord in the track. Often that's quite straightforward... if you can loop the track, keep hitting keys until you find a note that matches the track best. If the bassline's simple, it might just be the note the bassline's playing at any point in time.

Then, your first guess at a matching chord is... a simple major or minor triad in the key of the bass note you just identified. Most often, that'll be close; sometimes maybe the bassline's just moving through a note on its way to the root of a chord; sometimes you might want to try chords that match a 4th above or below the bass note.

You're looking to build up a chord progression that feels at home with the rest of the track.

Then drive it into your memory: practise it, play with transposing it, record it in a DAW, write it down somehow (the I / IV / vi notation style's useful cos it's not specific to a particular key).

You should be able to build up a bank of chord progressions, and if you're lucky you'll start to spot some repeating patterns - you'll start to see what chord progressions are common in your chosen genre.

Um... http://www.hooktheory.com/theorytab/common-chord-progressions

Then start taking your chords a little off the beaten track: adding extra 4ths, 7ths, 2nds/9ths... or trying out majors in place of minors and vice versa...

I'm always trying to create nice, meaningful chords when I try to produce trance. It rarely happens, however.

Are there any tricks that you guys use? Anything worth sharing?

Thanks!

What is the story you want to convey in your song?

You should be a little more specific than that! When I look at my archive of unused parts, I've labeled many of them with mood keywords, but I try to stay away from just calling them "emotional" because that's an empty label that says nothing about which emotion it actually is.

i've noticed that writing in the minor chord works for trance, its all about hitting that right note and you know it when you hear it, i dunno thats what I've noticed....music theory is really important but you have to go beyond that and think outside the box, its complicated haha...

The best advice I can give you is to become a student to the styles or artists that you like, and spend time learning their chord progressions.  Trance progressions are typically simple and repetitive, usually just using 4-5 chords (often with 7th and 9th extensions), and most often in pure minor keys.  I don't listen to a lot of trance anymore, but I rarely hear much that is written in major keys or that employ cadences (e.g., perfect, half, plagal, etc.) in the classical context - not that they're non-existent, but I just don't hear them very often.

Having some music theory knowledge is certainly helpful, but having a good ear and spending a lot of time trying to developing your ear training skills is probably a lot more useful than learning classical theory.  It seems like a lot of people put too much emphasis on theory, and fail to give equal time to the ear training and composition components (theory and composition are NOT the same thing), so it's important to approach it more holistically IMO.  Your ears should help you identify the tonic chord, whether it's major, minor, or modal, and intervals (focusing on the bass notes is a big help here).  Theory can help you rule out or identify the non-harmonic notes (i.e., notes that are outside the key/mode of the song) and help you narrow down the list of possible notes and chords that are based on the scale degrees.  It probably sounds more complicated than it is, but with a basic understanding of scales/modes, how triads and extensions are built on the scale degrees, and consistent practice over a few months, you should be able to start figuring out the chords in almost any trance or pop within a listen or two.  And, when you do, you'll quickly start to realize how many tracks share the same core chord progressions, but with their own little variations in voicing, extensions, melodic relationships, etc.  That will give you a ton of starting points.  Where you go from there depends on your own tastes, experimenting, borrowing other ideas and making them your own, etc.