Any camera meter takes a reading the scene and assumes on average that 18% of the light is reflected so the meter is only accurate if the scene is average. If you have a scene with lots of dark areas the meter will over expose because it's trying to make the dark areas 18% and vice versa.
For typical situations it works well and the camera also evaluative metering that looks are sections of the scene and compares the results with presets in its memory so it can spot situations such as backlighting, etc.
An incident meter just measures how much light falls on the scene so if you take a reading in the same light as the subject the exposure will be bang on for that 18% (which might not be right for the subject but is a good starting point).
They are two completely different ways of metering a shot so it's a little difficult to compare but a light meter will be more consistent than the camera's meter.
Some light meters also offer reflected or spot metering which is much closer to the camera's method but without any evaluation and checking against the examples in memory. Likewise most cameras offer some kind of partial metering which works a bit more like a light meter.
I've not used a light meter for 15 years or so now, I generally shoot in manual mode and take a reading from the scene get that exposed correctly and then adjust as the light/subject requires it.
An old technique is to take a reading off the palm of your hand and work out how much you need to under or over expose it. Then no matter where you are you can take a reading off your hand and get consistent results so long as the reading is taken in the same light as the subject. No need for expensive meters or to trust the camera's evaluative skills
hahaha what a ****.