If you have just single supply, you can follow the method suggested by dnanar.
There is more straightforward method, is by putting a power probe (or plot W trace) in pspice. The power plotted is the instantanious power. You would need to take average of this to get averaged power. Also measure this for time range when the circuit is in steady state for a number of cycles (if it is switching ckt).
If you are using behavioral devices in your circuit, then power plotted would be of no significance, as it would not truly represent real-life scenario.