Your circuit is already biased via the 20µA current source. If this is a correct value depends on your process and its model files. You'd perhaps have to adjust the current value and/or some W/L ratios until all MOSFETs are in saturation. M6 & M7 don't necessarily have to operate in saturation (could operate in triode region).
i'm sorry, how could i get the ICMR value then put it into the Vin+, Vin- then get the DC bias?
i heard there should be an equation to calculate? i'm quite confused!
so the vin_min=Vov1+Vov5+Vov6; Vin_max=Vdd-Vov3+Vth3, then use Vov=(2*Ibias/miu*Coxide*(w/L)) to calculate it, then input the value into DC analysis, to check each MOSFET in saturation? right?