Usually, I have a very messy way of thinking - ideas pop up, variations unfold then I switch to another idea, etc.
Sometimes I focus on one line if the position is very tactical, or if there is some sacrifice coming, and then I check carefully the tree variation, but it's quite rare.
i'm trying to come up with a plan and make moves that help me get position i would like/would be comfortabe to play
i think about the move that looks best, and play it! it doesn't work very well, i need to try thinking more often...
My way of thinking is more of a muttle in looking at the board for all the variations I can find and seeing the benefits and detriments of each move. I have a difficult time seeing more then five moves ahead as the variations get to complicated and lost.
First, a word of warning about "Think Like A Grandmaster". This is a great book, but Kotov explains pure calculation based on candidate moves. He speaks very little if none about the reasons for the candidate moves. You could calculate your brains out about every single available move and get nowhere. You should have a plan and moves should be based on that. Check out Silman's books "How to Reassess your Chess" and "The Amateur's Mind". I think you'll like what you find.
As far as myself, after 8-10 opening moves I begin to try to develop a plan around some weakness or imbalance. I then try to have all my moves, including any tactics, revolve around that plan.
In one word.............Carefully :) Well at least I try! Basically I just keep working to whatever plan I have decided on, keeping an eye out for errors by the opponent that allow me to profit elsewhere.
I also agree with whoever said to dispense with Kotov's work and pick up Silman's. Silman's is much more applicable to those of us below master level.
For me i usually at the beginning of the game I give up to flow, well unless he makes an fatal error after that i try to think of a tactic that will probably leave him without a lot of choice or the only choice to follow my game, if he disobeys me hes dead 
Of course there's one error, i usually overlook a single move that can mess my game for good, which will leave me in a hard position :)
Mate, material, initiative. Figure attacks on king with an end of mate or winning material. Then start with attacks on queen with an eye to winning material or pushing more valuable pieces around with less valuable pieces/pawns. Repeat with rooks and minor pieces.
Remember to look for tactical themes, unprotected material, creative ways to counterattack against what your opponent is doing to you (e.g. he attacks a knight with a pawn, you find a way to attack his rook). Don't let him just push you around.
Well it s hard to describe a thinking pattern for me.But there s is some mental check list i m carefull with..such as searching for forks, skewers..pins.When an opponent makes a move try to see what he left behind.And then when position arises try to make some tactics on the opponent..then i enter in hicetnunc mode of "messy way of thinking".
I got to think the one advantage the human brain has over a computer is the ability to quickly assign relevance to data, to prioritize rapidly on the most obvious threat
And that works against me half the time in chess.
Safety first! When I'm playing well, I look at my opponents move first. Where can he get to from where he just moved? What lines/attacks did he open up? After determining that there is no immediate threat, then I'll move on to my move. Its alot simpler when a good plan forces all of his moves to be predictable, it saves my time and then I can think about my plan and next move on his clock.
At this stage of the discussion, I cannot resist quoting this funny sentence, in reference to Kotov's famous method :
"I don't think like a tree - do you think like a tree ?" - GM Anatoly Lein
hicetnunc
Interesting post you have here. Here are some of the ways I think:
1. TACTICS I'm always looking for 1-5 move combos that I can use against my opponent, and I watch out for the ones my opponent is trying to use against me.
2. I always try to take time to think about the game in the middle game. I make it a point not to blitz the middle game because I know it won't help me improve as a player.
3. In terms of general strategy, when there's not a tactical shot available, I'm mainly concerned with taking my pieces off the back ranks, moving them to the center, and defending my King.
Piece activity I believe is the most important thing about human chess. I think pawn structure is very important, but piece strength is more important. I hope this helps.
I'm still a beginner, having only been playing a few months, so my line of thinking is probably wrong, but I usually focus on two things:
1 - Piece coordination. I don't like having a piece stranded out in the middle of nowhere unless it had very good reason to get there.
2 - Avoiding weaknesses. I usually play very defensive, whether I'm white or black, because (in close conjunction with point 1) I'm usually trying to avoid weaknesses that lead to blunders that lead to losses.
I've found that concentrating on these two things serves me pretty well. Against lower/similar ranked players I can usually develop my pieces nicely this way until my opponenet makes a mistake. Against higher ranked players I have difficulty once an all-out attack is made because I don't have any plan for counter-attack, so that should probably be something I'm thinking about as well.
Just figured I'd throw in a beginner's thought process...
I just want to point out that Kotov's recommended analysis tree is NOT how computers work. Kotov recommends searching each candidate move to full depth before moving on to the next candidate move. Computers do an iterative deepening. This means they search every move 1 ply, then order to moves best to worst and search again 2 ply, reorder and search again. This allows the computer to more efficiently prune the tree.
Humans generally do something similar. While calculating the consequences of one move, you'll get ideas about the position that apply to other moves. You might, for example decide to calculate what happens with a different move order.
I have read a few pages in the book "Think Like a Grand Master" by Kotov. In the book he teaches you how to think OTB and that is by collecting first all the candidate moves and then later think deep into each moves, variations by variations - like what a computer does (?)
My way of thinking OTB is similar but a little different because I sometimes consider intuition.
What about you? How do you think OTB?