This is how I understand them.
- Watch sb/sth doing sth: implies watching something continuing for a duration, often continuous or repetitive.
- Watch sb/sth do sth: implies the action watched is transient and short.
However, the distinction is not always clear cut, so the two may be used interchangeably in many cases.
Examples:
- I watched the car approach and zip by at a hight speed - the action is transient and one time occurrence.
- I watched her reading a book in her armchair - the action goes on over a period of time without much change.