I have an application which needs achieve phase shifted full bridge working at different switch frequencies, for eample, switch frequency is 300KHz ,500KHz and 700KHz, for me one frequency is OK, but make three switch frequencies achieve ZVS for one full bridge, it seems not easy, does anybody have good idea? Thank you!

It should be doable, it is not such a wide range of frequencies. That doesn't mean its easy, of course, no bridges are easy, and this adds a new twist to it.

You need to design transformer to not saturate at the lowest freqency, of course.

You don't mention the power level.

The duty cycle loss due to leakage inductance and/or external inductor at the lowest switching frequency of operation should not result in a lower duty cycle than required maximum duty cycle at the highest switching frequency.

I agree with Ray - it is absolutely doable. The question is always why phase shift..? Is it legacy product or new design. There are other soft switching topologies capable of the same. Considering the input voltage range (you did not mention it) and load range (also unknown) the task in hands could be not simple. So, please, talk to you company to consider a consultant at least for "jump start" period.

Previously we have a ZVS full bridge design with only one switch frequency, it works OK, but the new design we need this full bridge converter with ZVS under different switch frequencies, as I know an small inductor should be added in primary side to make it resonant with MOSFET output capacitance to make ZVS working, but this small inductor has relation with switch frequency, that means if I make it works under lowest frequency, all other frequencies maybe can not achieve perfect ZVS, this is the major concern for me.

Actually, the primary inductor must be sized for achieving full power at the highest frequency. At lower frequencies you will regulate by shifting the phase. Please, remember that Phase Shift PWM in it's raw form does not provide soft switching in full range of loads/bus voltages. If you have PFC in front of phase-shift PWM - it helps. The payment for universality is high peak currents at low switching frequency but... well that is how you pay for universality.

I would second Cong and others: freq range in not a problem per se. If ZVS is a paramount then one should look for auxillary circuitry which could easily double magnetics etc.