A pointing model is essentially a map of the telescope's pointing ability around the sky.  Even if the telescope is aligned properly, the weight distribution of the camera on the mount, the drag from the camera cables, and any other imperfections in the mount motors or positioning will cause the mount to have pointing errors.  A pointing model maps those imperfections in the pointing and helps compensate for them, which impoves the pointing accuracy.

The telescope control software for our Mount, TheSky6, works quite well with the pointing model application TPoint.  Using TPoint, we are able to create a pointing model quite smoothly by slewing the telescope to a number of bright stars in the sky, taking quick exposures with our camera, and checking that they are centered in the images.  We mostly use stars of magnitudes 2-5, with exposure times of 5 seconds.  By taking 1/8 subframes of our full field, we dramatically shorten the readout time and make it easier to check that the stars are centered.  We use about 25 points for our model, which is enough for the pointing accuracy that we want of less than 20 pixels, which is about 7 arcminues, or one half of one percent of our field width.