Hi Qingwen,
There is nothing unusual about your current waveform that I can see. The weirdness comes from the fact that the waveform is incomplete and hence does not decay to zero at the final time. This messes up the integration from psi4 to h and causes the strange bouncing around that you currently see. If you were to compare the raw psi4 data of the two waveforms, I'd expect they would agree quite closely. In general, you should probably refrain from looking at the strain data until your simulation has run to completion.
This issue is quite common with neutron star waveforms which don't always go to zero at the final time. It can be resolved by tapering the end of the psi4 data to ensure that it does decay to zero at the final time.
Best,
Shawn