Which one is better？

When designing an opamp, razavi gives his way in his book by starting from power budget and then allocation different Vdsat to each Transistor to set the W/L ratios ( see design analog cmos integated circuits) . While sansen, seemed from another point of view, designs the opamp by focusing on GBW and set Cc at the very begining, he also set the same Vdsat to be 200mV to each transistor ( see analog design essential). Could you tell me which method is better?

Depends on what's the need.

Traditionally, when designing discrete devices, we absolutely use razavi's method,it is effective and simple, but when we faced with IC, took opamp as an example, sometimes, a power constrain was not given as a spec.so it confused me...