I have had horrible times looking at datasheets for fuses, and their ratings are essentially very vague. The rating might be for 5A for example, but when you look at the datasheet it says that at 5A it might take 60 seconds to blow! But at 6A, the 5A fuse might blow in only 10 seconds, and at 7-8A the fuse might blow in 0.5 seconds etc. So if your servo operates at 500mA normally, but stalls at 2A, I would get a 1A fuse. It would only help for a prolonged stall condition.
The next thing to look at is using a quick blow fuse as you said, which are supposed to trigger really fast! The issue is when the servo starts, or is oscillating around a setpoint, or is changing positions often, the "stall" current occurs quite frequently - but not for long. How long? Who knows! It's very variable, very imprecise, and your quickblow fuse will be extremely annoying especially if it's not a resettable one, and even then it can take a while to cool down and reset.
I suggest finding a fuse which is "rated" for just above your normal operating current, so that many minutes of "average" current does not trip the fuse - but a few seconds worth of stall current will trip the fuse.
As I mentioned, find something which is rated to turn on just above (like 20-25%) the operating current - depending on the component and it's datasheet, this may only mean within a few minutes it will trip at this point. Then as the stall current remains in effect for many seconds, it should be 200% or more of your rated fuse trip current, meaning it should trip within a reasonable time span to not trigger during start-up current or fast changing of position/direction scenarios.