The traditional fault diagnosis methods for broadcasting and television transmitters have small changes in signal amplitude and low diagnostic accuracy. In order to solve this problem, a method based on VMD and wavelet transform is proposed to extract signal information, and the noise level of the decomposed signal is quantified by permutation entropy.The experimental results show that when the detection time reaches 4s, the time-domain waveform of the signal is complex, the overall amplitude increases significantly, and the amplitude is between [-10,10], so that the fault is identified here; when the signal to noise ratio is within the range of [-2,6], the diagnostic accuracy of different samples is between 98% and 99%, which can more accurately diagnose the fault in the transmitter.