Abstract: In the airborne sound transmission path analysis experiment, an external sound source is used to simulate car intake and exhaust noise. However, the existing low-frequency sound source cannot emit the 30 Hz low-frequency noise such as intake and exhaust, and the volume of the low-frequency sound source is too large, so that a transition pipe is needed to transmit the emitted noise signal to the exhaust tail pipe. In order to solve the problem of acoustic impedance changes caused by introducing transition pipe into the sound source system, an adaptive pre-filtering noise correction method is proposed. This method uses the minimum mean square error adaptive algorithm to fit the inverse transfer function of the system to obtain the best filter coefficients. The finite impulse response filter is constructed to correct the signal distortion of the combined sound source system. The simulation results show that, compared with the Wiener filter, this method realizes a signal amplitude fluctuation of ±5 dB in the range of 30~1 000 Hz, and the signal output of the nozzle is stable. The filter is built and tested based on NI Compact RIO. The tested result of noise spectrum curve is in good agreement with the simulated one, which proves the effectiveness of the method.