I am using SGP4 propagator in STK to propagate my GEO satellite. SGP4 need TLE as an input. In TLE, I found the inclination of my satellite is 0.048. But the inclination result of SGP4 is 0.15. The different is big enough. Can anyone help.me to explain this? Thank you.
As far as I can tell, this is nothing to worry.
What goes into a TLE is a specific kind of average elements, which are meant to simplify short-term propagations. Here "mean" does not imply a time-averaged value, but rather the result of some specific mathematical transformations. This paper explains it fairly well, but notice it uses a theory different than that of SPG4.
When you look at the output of SGP4, you actually get a position and velocity vector, which you've probably converted to osculating Kepler elements. You can check the equation that converts mean and osculating elements form the aforementioned paper to check if the discrepancy you've found is to be expected.
Finally, you might need to check that TLEs use the TEME reference frame, while you might be working with J2000, GCRF or similar.