do the gain waveform correct?but shouldn't it has a -40db drop after the -20db drop,y the curve so smooth?

 

Since it is a second order low pass filter you would expect a 40dB/decade roll-off once you get a reasonable way beyond the cut-off frequency. So your simulation is probably OK.

I'm not real sure about the -20dB drop issue you have raised - perhaps you need to clarify your question.

Also keep in mind this type of opamp filter doesn't continue to follow an 'ideal' second order response indefinitely with increasing frequency.

 

ya,i also want to ask why the filter does not follow ideal response at high frequency.

 

You have simulated your circuit with 741 opamp model. This opamp has poor performance and this is reflected in the model. Use an "ideal" opamp model if you want the ideal response.
Also I think then you are talking about the -20dB drop, you mean the -3 dB drop. But remember one thing. The -40 dB roll-off is the bode plot approximation. Even if this a very good approximation, it is not 100% accurate. And it certainly do not take components specifications into consideration.