If you are saying that this error relates to the particular measurement range being used on the instrument, then the error is 1.5% of the upper value in this range. (With negative ranges taking the more negative value.)
So if you mean the maximum absolute error in a reading then it would be the error associated with the maximum value (positive or negative) that you can read with the instrument on a particular range setting.
In that case if the max value you can measure on a range is 300 deg C the error associated with that is ± 4.5 deg C which represents 1.5% of 300
For 130 degs it will depend on what scale range you measure that on.
If it was measured on, say, a 100 to 200 deg range setting, then the max error would be 3 deg C, being 1.5% of 200
Edit: It could also possibly mean 1.5% of the actual range. In that case it would be 1.5% of 100 as the range is 100 degs from 100 to 200.
Sorry I'm a bit vague.
It would help to have more information about the instrument and what you are measuring.
It's a bit difficult without seeing or using the instrument.
Have for the reply, I have been informed that the Max expected error would be + or - 6 degree C.
Condidering the range -100 to 300, will be a total range of 400
+ or - 1.5% of 400 is + or -6 degree C.
Not sure about what the max expected error at 130 degree C for the same ranges.
That makes sense. If there is just one range from -100 to 300 then you have a total of 400 and a % uncertainty of ±1.5% of 400 giving ±6 degs.
Regarding the value for 130 degs it looks to me as though the uncertainty is still ±6 degs as it seems to be saying that the uncertainty is 1.5% of the total 400 deg range.
Normally % uncertainties are related to the actual measurements, which would give 1.5% of 130 here. (±2 degs).
However, taking this literally would mean you have an error of zero for a reading of zero degrees. This doesn't make sense.
In which case I would stick to the original idea that the absolute error is always 1.5% of the range. Here ±6degs as discussed.
If this turns out not to be the case I would be interested in knowing exactly what they mean here for future reference.