Massive data transmission tasks make the failure rate of radio and television transmission equipment continue to rise, resulting in the decline of the accuracy and efficiency of equipment failure state recognition. Therefore, a new method of radio and television transmission equipment failure state recognition based on dynamic fault tree is designed. Firstly, the vibration signal of radio and television transmission equipment is decomposed by wavelet packet to obtain the characteristic data of equipment running state; Secondly, the supervised discriminant projection manifold learning method is used to reduce the dimension of the obtained equipment operation state characteristic data; Finally, combined with the fuzzy set to the dynamic fault tree, the fault state identification of the radio and television transmission equipment is completed according to the feature data after the dimension reduction processing. The experimental results show that the proposed method has high accuracy, high efficiency, good overall recognition effect and high practical application value.