• First, I take a preprinted form (produced by OpenOffice.org writer) which contains a table with entries filled in and blank entries (you can download it from this site--it contains no viruses, but ALWAYS virus-scan documents from unknown sources!).  Filled-in entries include name, location (including coordinates), and instrument used.  Blank entries are for date and time, seeing, transparency, object name, constellation, and dec/RA coordinates of the object.  I fill this out before beginning.  This is about the time (when I am observing from home) that I realize I forgot to put a watch on and I go back inside to get it.  The form also has a 4-inch circle, containing 3-inch, 2-inch, and 1-inch concentric circles (in a bulls-eye pattern) as well as four diameter lines to divide the circle into eighths.  This helps me plot the stars with some accuracy.  I write down the time (precise to the minute) I start sketching
• I figure out which way, in the field, (celestial) north and east are, and mark those on the bulls eye on the form.  Specifically, north is always toward Polaris (if I am viewing Polaris, all directions are south!).  East is 90 degrees counterclockwise from north in binoculars or in the finder, or in the telescope without the star diagonal, and east is 90 degrees clockwise from north in the telescope with the star diagonal.  I must keep in mind that the telescope without the diagonal and the finder will invert the image--so I nudge the telescope toward polaris while looking at the field to be sure I get the right direction for north.
• I then take a look at the object and plot bright stars near the edge of the field.  Sometimes I center the object, sometimes I look for an "interesting" field, sometimes I choose the field to have bright stars at the edge to make it easier to find the exact field again and again.  Plotting these edge stars first helps me find the field.
• Next, I fill in the brightest stars throughout the field, as accurately as possible, estimating the distance from center and angle as accurately as I can, and using the circles and diameters on the form to plot them in the right places.  The brightest stars I make as spikey disks (as that is how they generally look), dark with the pencil.
• Now, I plot the object of interest.  If it is nebulous (as are most deep-space objects), I smudge the pencil marks with my finger to simulate the fuzziness.  I look critically at the object, trying to note color, bright or dim spots, shapes, and so on.  Those things I cannot draw easily (such as color) I make note of.  The same is true of stars or anything else in the field, by the way.  If by bad choice of observing location a tree branch is in the field, I draw that too.
• Now, I draw asterisms.  Near each bright star, I draw the pattern of dimmer stars (and near the object of interest too).  I draw them relative to the bright stars instead of using the "distance from center" and "angle" method I had used for everything up till now.  The upside of this method is (despite inevitable, for me, inaccuracies) asterisms look right and I can recognize them later.  The downside is I might misjudge the size of an asterism and put a star in one and a star in an adjacent asterism too close together or too far apart.  I have not yet found a solution to this "painting myself in a corner" problem.  I try to make the size and darkness of the dots correspond with the brightness of the stars.
• Finally, I look again for very dim stars, stars so dim that I only see them with averted vision and perhaps even then, for a second out of many seconds when the atmosphere settles down for just a moment.  I usually insist on seeing a faint star more than once to be sure it is not just my imagination (I have plotted stars that turned out not to exist!).
• During this, I sometimes do other things.  For example, if I see a meteor or satellite pass through, I draw a line with an arrow through the field of view to represent its path.  I then write notes about it (brightness, apparent speed, exact time it passed through, and so on).  In the case of satellites, I look it up later (either via my Starry Night program or heavens above).
• I might even do a few experiments, such as (with the telescope) time how long it takes an object to cross the field of view, or try to estimate the angular size of an object or distance between two objects by timing how long it takes them to enter or exit the eyepiece field.
• After all is done, I write the ending time down.  Later, I go inside and redraw the sketch (I didn't always do that, but have learned it greatly improves the quality of the sketches) using colored pencils, charcoal, etc.  and tracing paper (to accurately transfer the image) to get a good "negative image" of the field.  I scan it, then invert the colors, and voila! it looks like a (crude) photograph!  I use Starry Night and the internet to label interesting features of what I sketched.  I type up the notes and add any other thoughts to make a complete online observation record.

So, this is my sketching method.  I am not an artist, but I am improving as I go along.  I still have a long way for improvement.