the behaviour of your shown curve seems to be plausible, according to CMRR frequency responses shown in the literature.
But without more information of your actual circuitry no one will be able to tell you if it is accurate for your investigated structure.
how have you performed your analysis until now?
CMRR is tested by tying both opamp inputs together and apply them to the same voltage. Ideally no output signal would result.
If you are investigating a dual supply structure, the only difference sould be the input voltage swing (negative and positive).
How this is performed in HSPICE in detail I do not know as I do not use HSPICE.